article
stringlengths 0
745k
| abstract
stringlengths 0
92.7k
|
|---|---|
there are many situations in physics where a continuous system obeying a particular set of equations at a macroscopic scale can be modeled by a discrete microscopic system obeying a very simple set of local rules .
for example , in equilibrium statistical mechanics , simple lattice models such as the ising model capture the behavior of generic classes of critical systems at large scales . another interesting class of discrete systems are lattice gas automata @xcite ; these models describe systems of particles moving about on a lattice , obeying simple collision rules which conserve quantities such as mass and momentum . in the macroscopic limit , these systems obey navier - stokes or other hydrodynamic equations of interest . in the quantum domain ,
there are also examples of discrete microscopic systems which capture interesting macroscopic behavior . lattice - gauge theories ( see for example creutz @xcite ) give an approach to studying the partition function and spectra of quantum field theories by mapping these theories to statistical mechanical ensembles .
there are , however , few discrete models for describing the dynamical evolution of quantum systems which preserve important features such as unitarity .
an example of a quantum system for which a unitary discrete model is known is the dirac equation describing a relativistic particle moving in one spatial dimension .
as shown by feynman @xcite , this system can be described by a simple microscopic model of a particle moving on a 1d lattice according to a simple local rule which essentially corresponds to a unitary form of random walk .
a straightforward attempt to realize a dirac equation in more than one spatial dimension as a form of unitary random walk can not succeed . by using operator splitting methods ,
however , it was shown by succi and benzi @xcite that a sequence of random moves along single axes , alternating with transformations which diagonalize each of the dirac matrices in turn , can give an analogous construction in higher dimensions . the discrete model for the ( 1 + 1)-dimensional dirac equation has been of renewed interest recently@xcite , due partly to the possibility of simulating such unitary microscopic discrete systems by quantum computers . in particular , it has recently been suggested@xcite that a simple quantum lattice model can be constructed which describes the motion of a system of many particles moving according to the one dimensional dirac equation . in this paper
we consider a class of models closely related to the 1d dirac lattice model , which give rise to a nonrelativistic single - particle schrdinger equation in an arbitrary number of dimensions . in these models ,
the time development rule is given by a single local , unitary transformation matrix .
thus , we are essentially considering the motion of a single particle under a unitary random walk process .
for this class of models we show that the macroscopic equation of motion satisfied by the wavefunction corresponding to particle density is the schrdinger equation .
we show that such nonrelativistic models can be constructed for an arbitrary number of spatial dimensions .
we also show that an arbitrary potential can easily be included in these models .
it is natural to generalize from the single particle models to a second quantized many - body system .
such a model could be implemented very efficiently on a quantum computer , so that the number of computational steps necessary to simulate a single time step would only depend upon the size of the lattice and would not depend upon the number of particles in the system being simulated .
thus , our results could be used to efficiently simulate an arbitrary nonrelativistic interacting quantum many - body system on a quantum computer exponentially faster than the same calculation could be performed on a classical computer .
such a system would be an ideal example of quantum computing , since the computing elements could be built from a system of spin-@xmath0 particles on a lattice obeying simple local unitary time evolution rules .
the principal difference between our models and the 1d dirac lattice model ( and its generalization by succi and benzi @xcite ) is that in the dirac model , the unitary evolution rule satisfied by the wave function or particle at each time step is infinitesimally close to the identity transformation . as the lattice spacing @xmath1 goes to 0 ,
the unitary transition matrix is of the form @xmath2 , where @xmath3 is hermitian . in our models ,
we take the transition matrix to be independent of the lattice spacing .
this form of a time development equation makes the system nonrelativistic , but allows for a formulation in an arbitrary number of dimensions .
a closely related model was considered recently in one spatial dimension @xcite , where simulations were shown to be consistent with emergent behavior corresponding to a schrdinger equation . in this paper
we prove that this is the general behavior of such models , giving a simple algebraic relation between the transition matrix and the mass of the nonrelativistic particle .
we develop such models for the schrdinger equation in an arbitrary number of dimensions . in the first part of this paper we will consider lattice models for single - particle motion .
these models are essentially unitary lattice - boltzmann models@xcite . in the latter part of the paper we generalize to many - body systems and
discuss how a lattice of quantum computing elements could be used to describe the motion of a large number of nonrelativistic quantum particles .
we conclude with a simple numerical check of the analytic description of a sample model . as a simple example of the type of system considered in this paper , consider the lattice - boltzmann model with a configuration space defined by two complex fields , @xmath4 and @xmath5 , taking independent values on a lattice with one spatial dimension @xmath6 and one temporal dimension @xmath7 .
define the dynamics of this model to obey the equations @xmath8\ ] ] @xmath9.\ ] ] these equations give a unitary time evolution to @xmath10 . to understand how @xmath10 evolves in a continuum limit
, we can expand the equations of motion through 4 time steps , giving for example @xmath11\\ & & \hspace{0.3 in } + \frac{i}{4 } \left [ \psi_2 ( x-2,t ) -\psi_2 ( x , t ) - \psi_2 ( x+2,t ) + \psi_2 ( x+4,t ) \right].\end{aligned}\ ] ] taking a continuous limit as the lattice spacing scales as @xmath1 in the @xmath6 direction and @xmath12 in the @xmath7 direction , we find the differential equation @xmath13 a similar equation holds for @xmath14 , and so it follows that @xmath15 thus , we see that the total amplitude @xmath16 satisfies a schrdinger equation .
as we shall demonstrate , this is the generic behavior of a unitary boltzmann model with a fixed time development rule .
we introduce the schrdinger model in sec . [
sec : se1d ] by presenting the one - dimensional case .
the model is generalized to cartesian lattices of arbitrary dimension in sec .
[ sec : send ] ; in this section we also discuss the inclusion of a potential . in sec .
[ sec : gas ] , we discuss how the one - particle models can be generalized to construct a quantum lattice - gas model of many nonrelativistic particles . in sec .
[ sec : numeric ] we give the results of a simulation of a single free nonrelativistic particle in 2d , comparing numerical results with the theoretical framework presented here .
the appendices give explicit formulas for models in 2d and 3d on a cartesian lattice .
in this section , we consider unitary lattice - boltzmann models describing the evolution of a single particle in one dimension .
keeping the collision operator fixed in the scaling limit , we show that a very general class of microscopic models give rise in the continuum limit to a schrdinger equation .
we define the model on a lattice given by points @xmath17 where @xmath18 is an integer .
the lattice can be taken to either have periodic boundary conditions or to be of infinite extent .
the state of the system at a fixed value of the time parameter @xmath7 is described by a wave function @xmath19 which depends upon the discrete position @xmath6 and an `` internal '' index @xmath20 taking values from 1 to @xmath21 , labeling possible particle velocities at the lattice site @xmath6 .
as in lattice - boltzmann models , at each discrete time step the various components of the field at each site undergo a local unitary `` collision , '' and then the @xmath22th component of @xmath23 propagates along the @xmath22th lattice vector @xmath24 to the new site @xmath25 to yield the new state of the system at time @xmath26 .
we consider only linear processes , so this interaction can be specified by an @xmath21-by-@xmath21 scattering matrix @xmath27 .
we take the continuum limit of the theory by scaling @xmath28 , where @xmath29 . in this limit
we will find that the discrete equation describing the dynamics of @xmath10 becomes a continuous differential equation , which we identify as the schrdinger equation .
in this section we will assume that each lattice site has two associated possible particle velocities , corresponding to right- and left - moving particles .
we will also assume that the dynamics is symmetric under right - left reflection .
more general models can easily be analyzed using a similar formalism .
the equation of motion for the model reads @xmath30 where @xmath31 correspond to right and left moving particles , so that @xmath32 .
the quantum wave function @xmath19 is normalized so that @xmath33 for all @xmath7 .
the matrix @xmath34 is a two - by - two matrix which is unitary so as to preserve the condition eq .
( [ eq : normalization ] ) .
we begin our analysis by transforming to the wavefunction @xmath35 where @xmath36 .
we can then expand the difference @xmath37 in the infinitesimal parameter @xmath1 , to get @xmath38 where @xmath39 is the two - by - two matrix @xmath40 because we are assuming that the interaction described by the matrix @xmath27 is invariant under reflection , @xmath27 must be of the form @xmath41 where @xmath42 and @xmath43 are complex numbers . because of unitarity we have @xmath44 .
@xmath27 can be put in diagonal form by writing @xmath45 where @xmath46 we can redefine @xmath27 and @xmath10 up to a phase , so that without loss of generality we can take @xmath47 where @xmath48 is a complex number with magnitude 1 ( @xmath49 ) .
we then have @xmath50 if we write @xmath51 , then we have @xmath52 at this point we would like to scale the time step as a power of @xmath1 so that this equation can be written as a differential equation in time .
however , there is a difficulty which arises due to the fact that there are two relevant time scales involved in the dynamics of @xmath53 .
there is an order-@xmath1 change to @xmath53 at every time step ; however , this order-@xmath1 term has a phase angle which rotates at every time step .
thus , the order-@xmath1 dynamics average out after a large number of time steps , so that the time - averaged rate of change of @xmath53 actually goes as @xmath12 .
the dynamics we are interested in are independent of the short - term order-@xmath1 fluctuations , so we must perform another transformation to remove these effects . with this goal in mind , we write @xmath54 where @xmath55 this equation is solved by @xmath56 where @xmath57 this can be solved for @xmath58 as long as the only nonzero entries @xmath59 appear where the @xmath60 and @xmath22 eigenvalues of @xmath61 are different .
we can now write a final dynamical equation for @xmath62
. @xmath63 if we assume that the unit of time scales as @xmath12 , we have the continuous dynamical equation @xmath64 we can now substitute the known matrices @xmath65 to compute @xmath66 @xmath67 @xmath68 using these matrices we have @xmath69 writing @xmath70 , we have @xmath71 thus , the dynamical equation for @xmath62 becomes @xmath72 where @xmath73 the equation for the first component of @xmath62 is thus precisely a schrdinger equation for a particle moving in one dimension with mass @xmath21 . to leading order
, @xmath74 is related to @xmath10 through the sequence of transformations described above , so that @xmath75 the first component of @xmath74 is therefore given by @xmath76 and this satisfies the schrdinger equation in the continuum limit , @xmath77 note that by taking @xmath78 we get @xmath79 , giving precisely the example discussed in sec . [ sec : intro ] .
we shall demonstrate in sec .
[ sec : send ] that , in an analogous fashion , in higher - dimensional theories the sum of the wave function components forms a scalar quantity which satisfies a schrdinger equation .
in this section we derive the general form for the continuum limit of the dynamics for a unitary lattice - boltzmann model with fixed collision matrix on a lattice with any number of dimensions . specializing to the case where the lattice is cartesian and
the collision rule is invariant under discrete rotations , we find that a generic collision rule gives a schrdinger equation in any dimension @xmath81 .
the analysis of the continuous equations of motion in @xmath81 dimensions proceeds in a fashion very similar to the discussion in the previous section .
we assume that the lattice contains a set of points @xmath82 , and that at each lattice site there are particle velocities labeled by @xmath20 , corresponding to velocity vectors @xmath83 in the lattice .
denoting spatial indices by @xmath84 , we denote the @xmath84th component of the @xmath20th velocity vector by @xmath85 .
the dynamics of the lattice - boltzmann model are described by the equation of motion @xmath86 where @xmath27 is unitary . transforming as before @xmath87 we have @xmath88 where the diagonal matrices @xmath89 are given by @xmath90 with @xmath91 being the @xmath84th spatial component of the @xmath22th lattice vector .
writing @xmath92 , @xmath93 we have @xmath94 we write @xmath54 where @xmath95 this is solved , as before , by @xmath96 where @xmath97 again , this can be solved for @xmath58 as long as the only nonzero entries @xmath59 appear where the @xmath60 and @xmath22 eigenvalues of @xmath61 are different .
the resulting continuum equation for @xmath53 is @xmath98 this is the general form of the dynamical equation for a unitary lattice - boltzmann model .
we now specialize to the case where the lattice is cartesian , so that there are @xmath99 possible particle velocities at each lattice site , corresponding to vectors of magnitude @xmath100 in each of the @xmath81 directions .
we choose the collision matrix @xmath27 to be invariant under the symmetry group of the lattice .
we will show that generically the continuum limit of the equation of motion is a schrdinger equation , just as we found for a general collision matrix in 1d on the cartesian lattice .
the constraint that @xmath27 is invariant under discrete rotations and reflections is actually quite a strong condition .
the @xmath99-dimensional space of velocity vectors transforms under a linear representation of this discrete group .
this representation contains only 3 irreducible representations , which allows us to determine @xmath61 up to 3 distinct eigenvalues . because of the symmetry constraint , we can always diagonalize @xmath27 by the matrix @xmath101 where @xmath102 .
the rows of this matrix consist of the 3 groups of vectors in the irreducible representations of the rotation group mentioned above .
the first row is the normalized vector @xmath103 .
the next @xmath81 rows are normalized versions of the vectors @xmath104 with @xmath105 in position @xmath60 and @xmath106 in position @xmath107 .
the last @xmath108 rows are vectors with equal components @xmath60 and @xmath109 , subject to the condition that the sum of the components vanishes .
this matrix puts @xmath27 in the diagonal form @xmath110 where the eigenvalue @xmath111 appears @xmath81 times and the eigenvalue @xmath112 appears @xmath108 times . by a simple phase redefinition we can choose @xmath113 .
we will furthermore set @xmath114 , which as we shall see will give rise to a schrdinger equation in the continuum limit . as we shall discuss later , however
, any value of @xmath115 gives a schrdinger equation ; we use the @xmath114 condition merely to simplify the presentation . with the stated conditions on @xmath61
, we can compute @xmath27 .
we find that all elements of @xmath27 are equal to @xmath116 , except the matrix elements @xmath117 and @xmath118 , which are equal to @xmath119 .
thus , @xmath120 at a microscopic level , this collision matrix gives an equal amplitude for a particle to move in every direction other than directly backwards . to check the unitarity condition
, we verify @xmath121 and @xmath122 we can now proceed to calculate the other matrices needed for the dynamics .
there are @xmath81 matrices @xmath123 . for a particular value of @xmath84
, we find that all matrix entries vanish except those in the @xmath124th row and the @xmath124th column , which are given by @xmath125 where by @xmath126 we denote a sequence of @xmath20 0 s , and where we have defined @xmath127 we can now immediately compute @xmath128 , which has nonzero elements in the same positions , given by @xmath129 @xmath130 we are now interested in computing the differential equation describing the continuum limit of the dynamics of the first component of @xmath62 , which we will denote by @xmath131 . as before , we define @xmath132 to compute the dynamics of @xmath131 , we need to know only the first rows of the matrices @xmath133 and @xmath134 . from the above expressions
, we find that the first row of @xmath135 is given by @xmath136 from the above form of @xmath137 , we see that the first row of @xmath134 is given by @xmath138 thus , the first row of the combined matrix is @xmath139 where @xmath140 with @xmath141 as a result , we obtain the differential equation describing the dynamical evolution of @xmath131 in the continuum limit , @xmath142 which we recognize as schrdinger evolution in @xmath81 dimensions . in appendices [ sec : a1 ] and
[ sec : a2 ] we work out the specific cases of @xmath143 and @xmath144 in detail .
note that had we chosen the eigenvalue @xmath112 differently , the difference would have appeared in the last @xmath108 rows and columns of @xmath58 .
following through the computation , we find that the only change would be that new terms would appear on the right hand side of eq .
( [ eq : schrodinger ] ) , proportional to the eigenvectors of @xmath27 with eigenvalue @xmath145 .
however , these terms would have had a phase @xmath146 , and thus would have averaged out in the continuum limit as long as @xmath147 , making no change to the final result eq .
( [ eq : schrodinger ] ) .
thus , we reach the conclusion that for any collision rule invariant under the lattice symmetry group , so long as @xmath48 is distinct from the other eigenvalues of the collision matrix , the resulting continuum dynamics for the total amplitude @xmath131 are governed by a schrdinger equation . in general , we can easily include an arbitrary potential @xmath148 by including a position dependent phase in the transition matrix @xmath27 .
if we perform the above analysis for a model with transition matrix @xmath149 where @xmath27 is a spatially invariant matrix such as discussed above , then the general form of the dynamical equation becomes @xmath150 this becomes for the total amplitude @xmath131 , the schrdinger equation in an external potential @xmath151
we now consider models in which multiple particles move independently according to the schrdinger equation in @xmath81 dimensions .
one way of simulating the motion of @xmath18 particles in @xmath81 dimensions is to introduce extra degrees of freedom for each particle .
thus , for example , we could model the motion of 2 particles in one dimension by the lattice - boltzmann model @xmath152 where @xmath153 are the positions of the two particles , @xmath154 are the internal indices specifying their directions , and @xmath27 is a 2-by-2 matrix for unitary schrdinger evolution in one dimension , as discussed in sec . [ sec : se1d ] .
notice that this dynamics is equivalent to that of a single particle moving in two dimensions . in a similar fashion
we can describe models where @xmath18 particles move in @xmath81 dimensions , by constructing a unitary lattice - boltzmann model in @xmath155 dimensions .
it is straightforward to incorporate an arbitrary interparticle potential in this formulation ; the potential is a function of the particle positions and can be included as discussed in section iii.c .
this gives a procedure for simulating an interacting nonrelativistic quantum many - body system on a classical computer .
although this may give a useful algorithm for systems containing only a few particles , if we wish to simulate the motion of a large number of particles using the method just described it is clear that the number of calculations needed to perform even one time step of the evolution become rapidly intractable .
for example , simulating the motion of @xmath156 particles in three dimensions on a lattice of side length @xmath157 would take on the order of @xmath158 calculations per time step , beyond the capacity of any imaginable classical computer . however , the technology of _ quantum computing_@xcite presents a paradigm in which such calculations can be done .
we will now describe a way in which the above algorithm can be implemented on a quantum computer with a speedup exponential in the number of particles .
in fact , it is natural to simultaneously perform the calculation for all numbers of particles which will fit on the lattice , essentially performing a discrete simulation of nonrelativistic quantum many - body theory .
the resulting model falls in the class of quantum lattice gas automata , which were recently defined by meyer @xcite in the context of the @xmath159-dimensional dirac model .
the exponential speedup of this algorithm on a quantum computer is a specific example of the general observation by feynman @xcite and lloyd @xcite that quantum mechanical systems can be simulated more efficiently on a quantum computer than on a classical computer .
a quantum - computing device is composed of simple quantum elements such as particles with spin @xmath0 ( q - bits ) .
the state space of the system at any fixed time is the tensor product of the hilbert spaces of the states of the elementary computational elements .
thus , for example , a system with @xmath21 q - bits has a state space of dimension @xmath160 . at each time step ,
some small number of q - bits ( usually two or three @xcite ) are subjected to a unitary time evolution , described by acting with a unitary matrix on the hilbert space of the affected elements .
quantum computers have recently become of great interest because of the result due to shor@xcite that it is possible to factor large integers on a quantum computer in polynomial time , a procedure thought to be impossible on a classical computer . in order to implement the many - body simulation described in the beginning of this section on a quantum computer , it is necessary to make some restrictions on the behavior of the many - body wavefunction under exchange of particles . the example system in eq .
( [ eq : boltzmann2 ] ) describes two particles moving in one dimension without interacting . in this model
, both particles can be moving in the same direction from the same lattice site at a given point in time .
we can modify this model slightly to give the particles exclusionary ( fermi ) statistics by making the transition matrix at @xmath161 force the two particles to move in different directions .
this corresponds to introducing a contact interaction between the two particles when they move within a single lattice distance . by making the initial conditions antisymmetric under exchange of @xmath6 and @xmath162
, we have a simulation of two nonrelativistic fermions moving in one dimension .
alternatively , we could symmetrize the wavefunction and we would have a simulation of `` hard bosons '' which obey bose statistics but which can not occupy the same lattice site .
either of these approaches naturally generalize to arbitrary numbers of particles and arbitrary dimensions .
for the remainder of the discussion we assume that the particles obey bose statistics .
the issue of implementing fermionic systems on a quantum computer is more subtle @xcite , and has recently been addressed by abrams and lloyd @xcite . in previous sections we discussed the motion of a single particle , with a wave function @xmath163 .
now , we would like to consider the state space for a quantum system of many particles .
a natural basis for the hilbert space of such a system is the set of states in the fermionic fock space associated with the spatial lattice ; such states are identified by a set of occupation numbers @xmath164 ( taking values 0 or 1 ) for each possible particle position @xmath82 and internal index @xmath20 .
the hilbert space of the model is thus @xmath165 dimensional , where @xmath21 is the number of possible values of the internal index , and @xmath166 is the number of lattice sites .
for example , a basis vector of the state space for a one - dimensional system with 4 lattice sites @xmath167 might be given by @xmath168 where each ordered pair corresponds to the occupation numbers at a given lattice site .
thus , this state corresponds to the configuration where a single particle is at @xmath169 , with @xmath170 , and both particle positions at @xmath171 are filled .
the state of the quantum system at any given value of the discrete time parameter @xmath7 is given by a vector @xmath172 where the sum is taken over all basis vectors of the hilbert space .
this state is defined by the coefficients @xmath173 . in the quantum computing paradigm
, this corresponds to the state space of @xmath174 independent q - bits .
we will now define a quantum lattice gas automaton by defining a dynamics on the quantum state space .
the dynamics of the quantum lattice - gas will be described in two steps , just as in classical lattice - gas automaton models .
first there is a collision step in which the particles at each lattice site interact .
then there is an advection step , describing the propagation of the particles in the directions associated with the vectors @xmath83 .
each of these steps is described in the quantum system by a unitary transfer matrix acting on the state space of the system .
the total dynamics can then be described by the equation @xmath175 the advection step simply corresponds to a permutation matrix @xmath176 on the basis vectors described above , where each bit is moved forward in the direction corresponding to the appropriate vector @xmath83 .
for example , acting on the state in eq .
( [ eq : simplestate ] ) , the result of applying the advection operator would be @xmath177 where we assume periodic boundary conditions on the lattice . the particle which was at lattice site @xmath169 has been advected to @xmath178 , and the two particles which were at @xmath171 have moved to @xmath179 and @xmath169 .
we now consider the collision part of the time development rule .
the collision process is defined by a single unitary @xmath160 by @xmath160 matrix @xmath180 , which acts separately on the quantum bits associated with each lattice site .
thus , the state of the system is transformed by the unitary matrix @xmath181 given by the @xmath166-fold tensor product of @xmath180 .
we would like the collision matrix @xmath180 to have the property that it conserves particle number .
thus , this matrix is block diagonal in the subspaces of the hilbert space at each lattice site corresponding to a fixed particle number .
we have now defined a discrete model for quantum many - particle systems . to understand the behavior of this model in the continuum limit ,
let us consider the behavior when the number of particles in the system is relatively small compared to the number of lattice sites . in this case , at most lattice sites the number of particles present will be either 0 or 1 .
this part of the dynamics , which describes the free propagation of single particles , is described by the part of @xmath180 in the single - particle hilbert space . however , because this is a unitary matrix , generically the dynamics described by this transition matrix is precisely that which we studied in the previous sections , and corresponds to a nonrelativistic particle propagating according to the schrdinger equation .
thus , for relatively sparse systems , this quantum lattice - gas model simulates a system of many nonrelativistic particles whose free propagation is given by the schrdinger equation .
the remaining parts of the collision matrix @xmath180 describe a contact interaction between the various particles .
let us now discuss the computational complexity of the quantum algorithm . to implement the advection transformation by using quantum computing elements , it is only necessary to perform a series of exchanges of the values of the quantum bits representing the particle occupation numbers .
the number of such exchanges is essentially equal to the number of bits , @xmath174 ( recall that on a euclidean lattice of dimension @xmath81 , @xmath182 , so that for example if @xmath183 , the advection operation can be implemented in approximately @xmath184 quantum operations ) .
the matrix @xmath180 acts on the hilbert space associated with a subset of @xmath21 of the q - bits in the system .
counting degrees of freedom , generically such a matrix can be implemented with approximately @xmath185 elementary quantum operations on pairs of q - bits .
for example , in a 3d system , it would take on the order of 300 quantum operations to implement each @xmath180 matrix , so that the number of computational steps needed to perform the transformation by @xmath186 would be around @xmath187 .
note , however , that the part of @xmath180 which acts on the multiple particle hilbert space simply changes the phases of a delta function type interaction between the particles .
since these phases may not affect the results in most problems of physical interest , these components of @xmath180 can be arbitrary .
thus , in practice we need only find a combination of operations on q - bits which will give a matrix @xmath180 which preserves particle number and gives the desired symmetry properties and eigenvalues , reducing the number of steps needed significantly below 300 .
combining these observations , we see that this model can be simulated with on the order of @xmath166 elementary quantum computations at each time step ( or on the order of 1 if we are using a quantum computing system which allows parallel computation ) . since this system automatically contains the multi - particle wave function for all possible particle numbers , we have achieved an exponential increase in speed over what was possible on a classical computer .
the system as defined so far only includes interactions between the particles in the form of delta function interactions parameterized by the components of @xmath180 in the multiple particle space .
we can introduce an arbitrary interparticle potential @xmath188 by hand , by multiplying the wave function at each time step by the tensor product over all pairs of q - bits @xmath189 where the matrix @xmath190 acts on the hilbert space associated with the q - bits @xmath191 and @xmath192 , changing the phase of the wavefunction only in the component where both q - bits have the value 1 . implementing this interparticle potential
will take on the order of @xmath193 quantum computations for each time step .
although this significantly increases the computational complexity of the quantum algorithm , this is still exponentially faster than the analogous classical algorithm , since the particle number @xmath18 does not affect the complexity .
note that , unlike the rest of the algorithm , the implementation of interparticle potentials involves nonlocal interactions on the lattice . to clarify the discussion
, we consider a simple example of a collision matrix @xmath180 . for a many - body system in one dimension
, at each lattice site the collision matrix @xmath180 is a 4-by-4 matrix , acting on the hilbert space with basis @xmath194 , @xmath195 , @xmath196 , @xmath197 . since we are assuming that particle number is conserved and that the dynamics is symmetric under left - right reflection , the matrix @xmath180 is of the form @xmath198 where @xmath199 are complex numbers satisfying @xmath200 . by a simple global phase redefinition , we can choose @xmath201 .
the part of @xmath180 in the single - particle hilbert space is precisely the form of the collision matrix @xmath27 from sec .
[ sec : se1d ] . from the eigenvalues of this matrix
we can determine the mass of the free particles in the model .
finally , there is a single parameter @xmath202 which describes the phase with which two particles `` bounce '' . in this simple one - dimensional model
, there is therefore little freedom in choosing the particle interaction . in higher dimensions
there would be nontrivial phases describing delta function interactions between up to @xmath99 particles .
one major concern in the implementation of any algorithm is the issue of precision .
this problem is particularly acute on a quantum computer , where each quantum operation involves acting on the state with a unitary transformation which can only be controlled up to some finite precision .
furthermore , on a quantum computer there is the related but distinct problem of decoherence which must be addressed in order for any quantum computation to be feasible .
there has been a great deal of work recently describing how these problems can be solved using dynamical quantum error correction methods @xcite . without going into this issue in depth
, we make the simple observation that even without error correction , if the precision of each quantum operation is @xmath203 , then at each time step the error in the wave function will take a random step in the hilbert space with size @xmath204 . only after on the order of @xmath205 operations
will this error become significant .
thus , if we could achieve a precision better than @xmath206 , we could perform @xmath207 quantum operations successfully , which would allow us to simulate for example an interacting 3d system on a lattice with size of order @xmath208 . with the error correction schemes described in @xcite , there is in principle no upper bound on how large a system could be simulated , other than the size of the quantum computer which could be built to perform the simulation .
finally , we consider the issue of measurement in quantum lattice gases . in a classical lattice gas ,
hydrodynamic quantities , such as mass and momentum density , are obtained by averaging particles mass and momentum over blocks in space and/or time . in a typical lattice - gas simulation , this is done from time to time to obtain the macroscopic variables of interest .
the process of measuring these quantities is purely passive that is , their measurement does not affect the subsequent dynamical evolution at all .
in contrast , the analogous operation for a quantum lattice gas would involve occasionally measuring the state of some subset of the q - bits in the system , thus collapsing the quantum wavefunction onto the eigenstates of the ( space and/or time ) block number operator .
the set of quantities which are accessible through this type of simulation are rather different from those accessible through simulation methods on a classical computer .
for example , the dynamics of the system defines an effective hamiltonian which is an approximation to the hamiltonian of the many - body quantum system being simulated , however the spectrum of this hamiltonian is not directly amenable to measurement . instead
, the types of observables which can be measured in the simulation are precisely equivalent to the types of observables which can be measured in an actual interacting quantum system . for example , a typical experiment might be to initialize the system in a particular known state at time @xmath209 , and to ask for the probability @xmath210 at time @xmath7 that there is a particle in a region of space @xmath211 .
just as in the physical quantum system , we can ask such a question of our simulation ; we can perform the experiment a number of times , and each time we will find a particle with probability @xmath210 . to actually compute @xmath210 to some degree of accuracy requires repeating the experiment a number of times .
to test the algorithm , we consider the dispersion relation of plane waves in periodic geometry in two dimensions .
we consider a periodic grid with dimensions @xmath212 by @xmath212 , and initialize it with a plane wave of the form @xmath213 for @xmath214 , where @xmath215 where @xmath216 and @xmath217 are integers . choosing units where the spatial dimensions are of unit length , we have @xmath218 , and @xmath219 we evolve this initial condition in time , using eq .
( [ eq : tdcm ] ) with @xmath220 ( hence @xmath221 ) for the collisions .
every four time steps , we measure the inner product of the wave function with its initial condition , @xmath222 the result should go like @xmath223 , so the ratio of two successive values of this quantity is @xmath224 and hence the frequency is given by @xmath225 for a given wavevector @xmath226 , we measure this frequency at many time steps @xmath7 and take an average .
we expect the evolution of the system to be governed by the schrdinger equation , @xmath227 since @xmath221 , this leads to the dispersion relation @xmath228 we performed a series of simulations , where we considered wavenumbers @xmath229 and @xmath230 , where @xmath231 .
the points plotted in fig . [ fig : disp ] show the measured frequency @xmath232 as a function of @xmath233 .
the solid curve is @xmath234 .
it is evident that the agreement is excellent in the `` hydrodynamic '' limit of small @xmath235 , but degrades due to lattice artifacts of order @xmath236 at higher wavevector magnitudes . to demonstrate this
, we include data for @xmath237 ( gray points ) and @xmath238 ( black points ) .
it is evident that the dispersion relation is valid for higher wavenumbers on the larger lattice .
we have considered a very general class of lattice models satisfying unitary time - evolution rules . we have shown that generic models of this type describe the evolution of a nonrelativistic particle according to the schrdinger equation in an arbitrary number of dimensions .
these models can naturally be used to construct quantum lattice gases describing nonrelativistic many - body physics in an arbitrary number of dimensions .
it is straightforward to include an arbitrary interparticle potential into these models .
there are many ways in which this work could be extended .
numerical simulations could be performed in an arbitrary number of dimensions with multiple particles and with nontrivial spatially dependent potentials , and the results checked in analytically tractable cases .
further analysis is needed to understand the behavior of the system in the regime where particles are dense .
it might also be interesting to consider more general collision rules which create and destroy particles , possibly including antiparticles with separate quantum numbers .
of course , the actual implementation of quantum lattice - gas models on quantum computing devices is something which may not be possible for many years , if ever .
however , these lattice models give a simple framework with which to study problems in many - body theory . furthermore ,
the methods described here are also quite practical for simulating systems of a few particles on a classical computer .
it may be that the exact unitarity of these models at a microscopic level will make them more stable and possibly more useful than currently used discrete methods such as finite - difference approaches .
we would like to acknowledge helpful conversations with francis alexander , peter coveney , eddie farhi , and jeffrey yepez .
bmb was supported in part by phillips laboratories and by the united states air force office of scientific research under grant number f49620 - 95 - 1 - 0285 .
wt was supported in part by the divisions of applied mathematics of the u.s .
department of energy ( doe ) under contracts de - fg02 - 88er25065 and de - fg02 - 88er25066 , in part by the u.s .
department of energy ( doe ) under cooperative agreement de - fc02 - 94er40818 , and in part by the national science foundation ( nsf ) under contract phy90 - 21984 .
999 = .6em frisch , u. , hasslacher , b. , pomeau , y. , _ phys .
* 56 * ( 1986 ) 1505 .
frisch , u. , dhumires , d. , hasslacher , b. , lallemand , p. , pomeau , y. , rivet , j .-
, _ complex systems _ * 1 * ( 1987 ) 648 - 707 .
wolfram , s. , _ j. stat .
* 45 * ( 1986 ) 471 .
creutz , m. , `` quarks , gluons , and lattices , '' cambridge u. press ( 1983 ) .
feynman , r.p .
, unpublished notes as reproduced in s. s. schweber , _ rev . mod .
phys . _ * 58 * 449 ( 1986 ) .
feynman , r.p .
, hibbs , a.r . , `` quantum mechanics and path integrals , '' mcgraw - hill , inc .
( 1965 ) pp .
succi , s. , benzi , r. , _ physica _ * d69 * ( 1993 ) 327 .
kauffman , l. , noyes , h. , `` discrete physics and the dirac equation '' , preprint slac - pub-7115 , march 1996 .
meyer , d. , `` from quantum cellular automata to quantum lattice gases , '' preprint quant - ph/9604003 , march 1996 .
succi , s. , `` numerical solution of the schroedinger equation using discrete kinetic theory '' , ibm ecsec preprint , 1995 .
benzi , r. , succi s. , vergassola , m. , _ phys .
reports _ * 222 * ( 1992 ) .
see , e.g. , @xcite , and deutsch , d. , _ proc .
roy . soc .
london _ * 400a * , 96 - 117 ( 1985 ) , and _ proc .
london _ * 425a * , 73 - 90 ( 1989 ) ; c. h. bennett , `` quantum information and computation , '' _ physics today _ , october 1995 .
feynman , r. , _ int .
j. theor .
_ * 21 * , 467 - 488 ( 1982 ) , and _ found
. phys . _ * 16 * , 507 - 531 ( 1986 ) ; lloyd , s. , _ science _ * 273 * ( 23 aug 1996 ) 1073 - 1078 .
shor , p. w. , `` algorithms for quantum computation : discrete log and factoring '' in _ proc .
35th annual symposium on the foundations of computer science _
goldwasser , s. , ieee computer society press , los alamitos , california ( 1994 ) 124 .
lloyd , s. , _ phys .
* 75 * ( 10 july 1995 ) ; barenco , a. , bennett , c.h . ,
cleve , r. , divincenzo , d. , margolus , n.h . , shor , p. , sleator , t. , smolin , j. , weinfurter , h. , _ phys .
a _ * 52 * ( 1995 ) 3457 - 3467 .
abrams , d.s .
, lloyd , s. , `` simulation of many - body fermi systems on a universal quantum computer , '' preprint .
shor , p. w. `` fault - tolerant quantum computation '' , at&t preprint quant - ph/9605011 ( 1996 ) , and references therein .
we now present the formalism described above explicitly in two dimensions .
the matrices @xmath61 and @xmath137 are given by @xmath239 @xmath240 this gives us the collision matrix @xmath241 we can compute @xmath242 @xmath243 and thus @xmath244 @xmath245 combining these matrices together we find @xmath246 with @xmath21 described as in eqs .
( [ eq : m ] ) and ( [ eq : m2 ] ) , with @xmath247 . as predicted by the general discussion above
, the total amplitude contained in the first component of @xmath62 satisfies a schrdingerequation @xmath248 where @xmath249.\ ] ] it is interesting to note that while the variation of the fourth component of @xmath62 contains an oscillating phase , and thus has no interesting behavior on the time scale of interest , the second and third components obey separate second - order differential equations analogous to the schrdinger equation , but without rotational invariance .
using the above formalism in 3d , we have @xmath250 @xmath251 @xmath252 @xmath253 @xmath254 from these matrices we find that the total amplitude @xmath255.\ ] ] satisfies the schrdinger equation @xmath256 where as usual @xmath21 is related to @xmath48 through eqs .
( [ eq : m ] ) and ( [ eq : m2 ] ) , with @xmath183 . | we consider a general class of discrete unitary dynamical models on the lattice . we show that generically such models give rise to a wavefunction satisfying a schrdinger equation in the continuum limit , in any number of dimensions .
there is a simple mathematical relationship between the mass of the schrdinger particle and the eigenvalues of a unitary matrix describing the local evolution of the model .
second quantized versions of these unitary models can be defined , describing in the continuum limit the evolution of a nonrelativistic quantum many - body theory .
an arbitrary potential is easily incorporated into these systems .
the models we describe fall in the class of quantum lattice gas automata , and can be implemented on a quantum computer with a speedup exponential in the number of particles in the system .
this gives an efficient algorithm for simulating general nonrelativistic interacting quantum many - body systems on a quantum computer .
8.5 in 6.5 in -.35 in |
the cms experiment at the large hadron collider ( lhc ) at cern ( geneva , switzerland ) will be equipped with a silicon pixel detector providing high resolution 3d coordinates of the tracks .
it will consist of three layers of pixel sensors ( 100 @xmath1 m @xmath2 150 @xmath1 m pitch size ) in the barrel region and two discs in the forward regions @xcite .
the innermost barrel layer is expected to be exposed to a fluence ) .
] of @xmath3 per year at full luminosity .
the second and third layer will be exposed to @xmath4 and @xmath5 per year , respectively .
the charge sharing induced by the lorentz force in the 4 t cms detector magnetic field can improve the spatial resolution .
hence it is very important to study the effects of charge collection efficiency and lorentz drift in irradiated silicon devices .
here we present measurements of lorentz angle and charge collection efficiency , and compare two sensor designs at different irradiation doses .
the measurements were performed in the h2 beam line of the cern sps in june and september 2003 using 150 - 225 gev pions .
a silicon reference telescope @xcite was used to allow a precise determination of the particle hit position in the pixel detector .
both , the telescope modules and the pixel front - end , were mounted onto a common frame ( see fig . [
fig : beam_telescope ] ) .
the beam telescope consists of 4 modules .
each consists of two 300 @xmath1 m thick single - sided silicon detectors ( @xmath6 mm@xmath7 ) with a strip pitch of 25 @xmath1 m ( readout pitch of 50 @xmath1 m ) , which are oriented perpendicularly .
the resulting intrinsic resolution of the beam telescope is around 1 @xmath1 m .
+ the pixel sensor with the readout chip is mounted on a rotating support positioned between the second and the third module .
a trigger signal is generated by a silicon pin diode .
the data acquisition system , including slow control ( temperature and bias voltage of the pixel sensor ) is written in labview and labwindows cvi ( national instruments ) running on a pc .
the analog signals are digitized in a vme based readout system by two caen ( v550 ) and one custom built fadcs @xcite .
the whole setup was placed in a open 3 t helmholtz magnet with magnetic field parallel to the beam .
the pixel sensors were cooled down up to @xmath8 c by means of water cooled peltier elements .
the cms pixel sensors are manufactured in the `` n - on - n '' technique , consisting of @xmath9 structures on n bulk silicon .
this allows a partially depleted operation of highly irradiated sensors after type inversion , but requires inter - pixel isolation .
two isolation techniques were considered in our latest prototype designs : p - spray @xcite , where a uniform medium dose of p - impurities covers the whole structured surface , and p - stop , where higher dose rings individually surround the @xmath10-implants .
[ fig : designs ] shows the layouts of these designs . due to possible failures in the pixel contacts ( bump - bonding )
an intrinsic biasing method has to be introduced .
this is realized by a bias grid and punch - trough structures @xcite for the p - spray detectors and by openings in the rings for the p - stop detectors .
all our test devices had @xmath11 pixels , with a sensitive area of @xmath12 mm@xmath7 , and a thickness of 280 @xmath1 m .
the readout pitch was 125 @xmath1 m @xmath13125 @xmath1 m .
+ some of the devices were irradiated in a 24 gev proton beam at the cern ps .
the sensors received total particle fluences of @xmath14 , @xmath15 and @xmath16 respectively . during the irradiation
the sensors were kept at room temperature and without bias , while after irradiation we stored and operated them at @xmath17 c. @xmath18-curves were measured @xcite before and after irradiation , all measured sensors showed a breakdown voltage above 600 v. finally the sensors were bump bonded to non irradiated readout chips of the type psi30/ac30 @xcite allowing for a non zero - suppressed analog readout of all 704 pixel cells . a fast ( @xmath19
ns ) external signal for the internal sample and hold mechanism had to be provided from the trigger pin diode .
the lorentz angle is obtained by the direct measurement of the charge drift in the magnetic field using the grazing angle method @xcite . the particle beam ( parallel to the magnetic field ) hits the pixel surface at a shallow angle @xmath20 ( see fig . [
fig : gr_angle ] ) .
the deposited charge drifts according to the combined electric and magnetic forces , resulting in a deflection of the particle track projection on the surface by the angle @xmath21 .
the lorentz angle is obtained from the relation @xmath22 . + the deflection measured at 3 t magnetic field is shown in fig .
[ fig : la_cluster ] for the non - irradiated p - spray sensor .
the angle @xmath21 is measured by slicing the histogram perpendicularly to the @xmath23-axis .
the position of the center of each slice is shown in fig .
[ fig : deflection_cluster ] as a function of the position along the @xmath23-axis . a measurement without magnetic field
is used to correct for the detector misalignment with respect to the beam ( bottom line in fig .
[ fig : deflection_cluster ] ) . for the irradiated device in fig .
[ fig : deflection_cluster ] there are two regions with a different slope , and hence , two different values of the lorentz angle . the difference in the lorentz angle can be explained by nonuniform distribution of the electric field in the irradiated devices .
different models of the electric field distribution in the irradiated silicon devices are discussed in refs .
@xcite .
the evidence of the non - uniform electric filed will be shown also in the next section . +
as most of the signal charge is collected from the region close to the pixel implant , this region was used to determine the lorentz angle .
figure [ fig : la ] shows the measured values for both non - irradiated and irradiated sensors extrapolated to 4 t magnetic field .
we observe a strong dependence of the lorentz angle on the bias voltage ( electric field ) , while it is only weakly affected by the irradiation or sensor design . for the non - irradiated sensors a lorentz angle of @xmath24 can be reached at a bias voltage of @xmath25 v , while irradiated sensors have to be operated at higher bias voltages , where the lorentz angle drops to roughly @xmath26 . since the electron mobility increases with decreasing temperature the lorentz angle measured at the lower temperature @xmath17
c is @xmath27 larger .
our values for the lorentz angle are in a good agreement with measurements and simulations in refs .
@xcite .
after irradiation the collected charge decreases due to charge trapping and partial depletion of the sensor .
the measurements of the charge collection efficiency as a function of the sensor depth were performed using the grazing angle method without magnetic field .
the averaged charge cluster profiles in non - irradiated and irradiated p - spray sensors are shown in fig .
[ fig : proj_cluster ] . in the unirradiated sensor
the charge is collected uniformly across the whole sensor depth , while the irradiated devices shows two regions with different collected charge . at low bias voltage in the irradiated detectors some charge
is also collected from the side opposite to the pixel implants .
this can indicate that the depletion starts from both sides of the detector , because of the nonuniform distribution of the electric filed .
the integral of the collected charge from the particle entry point up to a point along the @xmath23-coordinate is shown in fig .
[ fig : int_cluster ] .
there are two regions with different charge collection , represented by two different slopes of the curve for the irradiated device .
the total collected charge for different designs and irradiation doses as a function of bias voltage is shown in fig .
[ fig : q ] .
the charge collection depends on the position of the incident particle with respect to the pixel implant , caused by the regions with reduced sensitivity .
the average charge collected in the hit pixel as a function of the hit position is shown in fig .
[ fig : charge_dist ] for non - irradiated and irradiated sensors .
one can see that after irradiation the average pixel signal decreases , and the area with the reduced charge collection increases in the case of the p - stop design .
+ the signal - to - noise ratio ( s / n ) averaged over the whole pixel area is shown in fig .
[ fig : sn ] .
the measured signal - to - noise ratio is above 30 even after the irradiation .
we observe lower s / n ratio in the unirradiated p - stop device , because of the charge spread over several pixels caused by the p - stop openings . a more detailed study of charge collection for these sensors
can be found in ref .
two different designs of the cms prototype pixels sensors were tested up to 600 v bias voltage and after exposure to particle fluences up to @xmath16 . for the irradiated devices
we observe two regions across the sensor with different charge collection and value of the lorentz angle .
this result is an indication of a nonuniform electric field in the irradiated devices .
+ the lorentz angle mainly depends on the bias voltage and it shows a weak dependence on the irradiation dose or sensor design .
the lorentz angle at 4 t magnetic field reaches 26@xmath28 for the non - irradiated devices at a bias voltage of 100 v and it drops to 8.3@xmath28 for the ones irradiated at @xmath16 and at a bias voltage 600 v. + after irradiation and at high bias voltage ( 600 v ) the charge collection is about 60% of the value observed for the non - irradiated devices , and it is slightly larger for the p - spray design .
+ the signal - to - noise ratio decreases from 65 to about 35 after irradiation giving us confidence in operating the cms pixel detector up to the maximum expected irradiation dose . + for the irradiated devices the charge collection , and therefore the signal height , increases with the bias voltage but the lorentz angle decreases . for this reason the bias voltage must be optimized for the best performance of the cms pixel detector . +
we gratefully acknowledge silvan streuli from eth zurich and fredy glaus from psi for their immense effort on the bump bonding , maurice glaser and michael moll from cern for carrying out the irradiation , kurt bsiger from the zrich workshop for the mechanical construction , gyrgy bencze and pascal petiot from cern for the h2 beam line support and finally the whole cern - sps team .
the authors would like to thank dr .
danek kotlinski from psi for very useful discussions .
r. h. richter _ et al .
_ , nucl .
instr . and meth . in phys .
research a * 377 * ( 1996 ) 412 .
t. rohe _ et al .
_ , nucl .
instr . and meth . in phys .
research a * 409 * ( 1998 ) 224 .
t. rohe _ et al .
_ , ieee - nss , october 19 - 25 , 2003 , portland , oregon , usa , to be published . | the tracking system of the cms experiment , currently under construction at the large hadron collider ( lhc ) at cern ( geneva , switzerland ) , will include a silicon pixel detector providing three spacial measurements in its final configuration for tracks produced in high energy @xmath0 collisions . in this paper we present the results of test beam measurements performed at cern on irradiated silicon pixel sensors .
lorentz angle and charge collection efficiency were measured for two sensor designs and at various bias voltages .
, , , , , , , , , , , , , pixel ; radiation hardness ; lorentz angle ; charge collection ; cms ; lhc |
the phase behavior of fluids and fluid mixtures confined in porous and disordered materials represents a field of continuing theoretical and experimental interest due to a variety of applications in industrial technology @xcite . for example , porous materials are employed as adsorbents in many industrial processes , such as catalysis , adsorption separation , filtration and purification . experimental studies on the phase behavior of fluids adsorbed both in highly porous materials , such as silica aerogels @xcite , and in less porous ones , such as vycor glasses @xcite , have been performed .
all these studies have shown that the phase diagram of the fluids confined in the porous structures is strongly altered in comparison with the corresponding bulk systems . for @xmath0 @xcite and @xmath1
@xcite confined in very dilute aerogels it has been found that the liquid - vapor coexistence curve is much narrower and the critical temperature is lower than in the bulk .
it has also been observed a shift of the critical density towards the liquid phase and an increase of the coexistence vapor phase densities attributed to the attractive fluid - gel interactions .
very few theoretical studies on fluids confined in disordered porous materials have been performed . a theoretical approach based on considering the gel as a random field acting on the fluid succeeded in reproducing the important feature of spatial inhomogeneity @xcite .
important phenomena like wetting can not nonetheless be reproduced by such model . a model of `` quenched - annealed '' binary mixture was first studied by madden and glandt @xcite and then widely used in dealing with the problem of fluids phase separation in porous materials .
following this model , several integral equation theories and computer simulation studies have been successively proposed , in which the porous material is described by a random matrix of spheres @xcite .
in particular the simulation work done by page and monson @xcite dealt with the calculation of the phase diagram for a system representative of methane adsorbed in a silica xerogel and that of alvarez and al .
@xcite has investigated the sensitivity of the confined fluid phase behavior to the matrix realization .
the results of these studies about the vapor - liquid coexistence curve are in good qualitative agreement with experiments . the novelty of these works not observed in experiments is the occurrence of an additional phase transition between a medium density liquid and an high density one .
the liquid - liquid phase coexistence has been associated with the wetting properties of the fluid in the more dense regions of the adsorbent .
this second phase transition however has been found to be very sensitive to the matrix realization , while the liquid - vapor coexistence properties are seen to be robust to variations in the solid structure @xcite .
this behavior has been explained observing that the additional liquid - liquid transition is associated with the filling of low porosity regions in the matrix , which can occur for some configurations but not for others .
the picture which emerges from all these calculations is that the coexistence vapor and liquid phases for the confined fluids are disordered and inhomogeneous as a consequence of the adsorbent structure randomness .
the description of the aerogel as a random matrix of spheres however is not sufficiently realistic .
the gel is organized into a fractally correlated structure experimentally identified by a strong diffraction peak at small wave vectors measured by small angle x - ray or neutron scattering @xcite .
the fractal behavior of aerogels is associated with the irreversibility of the gel formation dynamics , which proceed by random colloidal aggregation of silica particles .
the fractal dimension of silica aerogels is about 1.80 in three dimensions @xcite
. structures with similar fractal dimension have been generated by monte carlo simulations using several hierarchical cluster - cluster algorithms @xcite .
in particular the diffusion limited cluster - cluster aggregation procedure ( dlca ) is a gel growth process widely used in theoretical studies of aerogels @xcite .
it has been shown that the structural properties and fractal dimension of dlca gels well agree with experimental data on silica aerogels .
a theoretical study of thermodynamical and structural properties of a lennard - jones fluid adsorbed in a highly porous dlca aerogel by means of an integral equation approach has been recently proposed @xcite .
the phase diagram as well as the structural correlations of the confined fluid are found to be influenced by the specific properties of the gel , such as its connectivity and fractal behavior , particularly at low fluid densities .
these results support the idea that a realistic modeling of the gel environment is necessary to deeply investigate the effects of the porous medium structural properties on the phase behavior of the adsorbate , especially on the second disorder - induced liquid - liquid transition .
in this paper we present the results of a grand canonical monte carlo simulation study about the phase behavior of a lennard - jones fluid confined in a dilute dlca aerogel , where fluid and gel particles have hard core diameters of equal size and interact by means of an hard sphere potential . in order to locate the liquid - vapor coexistence curve we have calculated the density distribution functions . the multicanonical ensemble sampling ( mes ) procedure has been employed to investigate thermodynamical states in the subcritical region @xcite . in the next section we describe the algorithm to build our confining system . in sec .
3 we give details of our simulation . in sec . 4 we present the results about the confined fluid phase diagram and compare it with that of the bulk system . sec . 5
is devoted to conclusions .
the algorithm we used to generate the aerogel configuration is the three dimensional off - lattice extension by hasmy and al .
@xcite of the dlca procedure first proposed by kolb and hermann @xcite .
the dlca process is an iterative method which starts with a collection of n identical spherical particles of diameter @xmath2 randomly placed in a cubic box of side l , with volume fraction @xmath3 .
aggregation proceeds via a diffusion motion of the particles .
if during their motion two clusters collide they stick together forming a new single aggregate .
the process is terminated when a single cluster is obtained .
periodic boundary conditions are used at the edges of the simulation box .
details of the dlca algorithm can be found in references @xcite .
we have generated with this algorithm configurations for aerogels containing 515 particles in a box of edge @xmath4 , corresponding to a volume fraction @xmath5 and a porosity p , defined as the ratio between the free space volume and the volume occupied by the gel , of 92% .
we have calculated for these aggregates the radial distribution function g(r ) and the static structure factor s(q ) following the lines of the simulation work by hasmy and al .
@xcite our results are reported in fig .
[ fig : aero ] .
[ fig : aero ] the radial distribution function g(r ) presents a strong peak at @xmath6 , associated with bonds between contacting particles , and other features corresponding to short range correlations between particles belonging to the same cluster ( see references @xcite for further details ) . before reaching the asymptotic limit of 1 , g(r )
exhibits a minimum corresponding to distances between particles located at the boundary of the clusters .
the location of this minimum gives an estimate of the mean clusters size @xmath7 . in our case
the mean cluster size is about @xmath8 . as a consequence of the presence of the minimum in g(r ) , the scattering function s(q ) , which is related to the fourier transform of @xmath9 , exhibits a pronounced peak at small q. the gel exhibits a fractal structure in the intermediate q - range .
the fractal dimension can be estimated from the power law behavior of the scattering function , @xmath10 in this q - range . in our case we have obtained @xmath11 consistent with the value expected for silica aerogels in three dimensions @xcite .
in our simulations the interactions between fluid particles are described by a lennard - jones potential with a collision diameter @xmath12 and a depth @xmath13 truncated at @xmath14 .
the aerogel particles are quenched and interact with the fluid particles by means of a purely repulsive hard sphere potential .
we assume that fluid particles have a collision diameter of the same size as the aerogel hard core @xmath15 .
the edge of our simulation box is @xmath16 . in the following lennard - jones units
will be used : @xmath12 for lengths,@xmath13 for energies and @xmath17 for temperatures . in order to study the phase diagram of the lennard - jones fluid
we have carried out monte carlo simulations in the grand canonical ensemble @xcite employing the algorithm used by wilding @xcite .
mc moves consist of either an insertion or a deletion attempt , proposed with equal probability .
particles movements are implicitly implemented through insertion and deletion moves . in order to calculate the liquid and vapor coexistence densities
the chemical potential can be varied , at fixed temperature , until a bimodal shape of the particles number distribution p(n ) is obtained .
the coexistence densities at the selected temperature correspond to the two peaks positions . to obtain also the location of the liquid - gas saturation line in the ( @xmath18,@xmath19 ) plane the equal peak weight criterion @xcite can be applied . according to this criterion
the chemical potential is tuned , at constant temperature , until the measured p(n ) is double - peaked with equal area under the two peaks .
the corresponding chemical potential belongs to the fluid coexistence curve .
this conventional grand canonical ensemble monte carlo technique becomes however impractical in the subcritical region we are interested in , due to the large free energy barrier separating the two phases and hindering spontaneous fluctuations of the system from the liquid to the gas phase and viceversa . in order to enhance crossing of this free energy barrier
we have used the recently proposed mes @xcite . with this powerful technique sampling
is made from a non boltzmann distribution , with a modified hamiltonian @xmath20 , where @xmath21 is the configurational hamiltonian .
the biased sampling function @xmath22 has to be chosen in such a way that the measured distribution @xmath23 is nearly flat . in this way
the mixed - phase configurations will be sampled with approximately the same probability as the gas and liquid configurations .
the best choice for the biased sampling function would be @xmath24 , where @xmath25 is the distribution we are looking for .
an estimate @xmath26 can be obtained by extrapolation making use of the histogram reweighting technique @xcite . in this approach
we can accumulate the joint probability distribution of system energy and particles number for a thermodynamical state characterized by a temperature @xmath27 and a chemical potential @xmath28 ( near the critical point where the interfacial tension is low ) @xcite : @xmath29 where @xmath30 is the density of states and @xmath31 is the grand partition function .
an estimate @xmath32 of this distribution for another thermodynamical state @xmath33 can now be provided : @xmath34 the new thermodynamical point @xmath33 has to be sufficiently close to @xmath35 in such a way that the statistical weight of the new configuration is not too different from the previous one . integrating over the system energy e , we can finally obtain a suitable bias function : @xmath36 at the end of the monte carlo simulation at @xmath33 with the biased hamiltonian we obtain the almost flat @xmath23 .
the real particles number distribution @xmath25 pertaining to @xmath33 is recovered from : @xmath37 .
this method has given more accurate results than the well known gibbs ensemble technique and gibbs - duhem integration scheme both for bulk and confined fluids @xcite .
we have performed extensive mes simulations in order to calculate the confined fluid phase diagram for a range of subcritical temperatures from @xmath38 to @xmath39 .
our simulations involved about @xmath40 steps for system equilibration and , depending on temperature , from @xmath41 to @xmath42 steps for calculating the ensemble averages .
runs of this order of magnitude are necessary in order to supply enough indipendent samples from each of the two phases and consequently obtain density distribution functions with enough accuracy for our analysis @xcite .
we note that on decreasing temperature the rate of the interchanges between the liquid and gas phases becomes more rare and longer runs are needed .
first of all we need to locate the coexistence curve of the confined fluid and in particular the near - critical region . for this sake we have calculated with standard grand canonical simulations adsorption isotherms of the density @xmath43 versus the chemical potential @xmath18 for temperatures ranging from @xmath44 to @xmath45 .
our results are shown in fig .
[ fig : isot ] .
figura2 [ fig : isot ] for a given temperature , we observe evidence of a coexistence at the same chemical potential between a dilute gas and a medium density liquid in our confined system . we have not observed in our isotherms the occurrence of a liquid - liquid phase coexistence , at variance with previous computer simulation studies on the phase behavior of lennard - jones fluids in disordered non fractal matrices of spheres @xcite .
in our calculations we have not considered desorption isotherms and the hysteresis associated usually with capillary condensation @xcite since we are interested in the coexistence between equilibrium phases .
we note that due to the slope of the liquid portion of the isotherms a small change of the chemical potential induces a strong variation in the liquid density .
therefore it is difficult to tune the chemical potential with the appropriate resolution to locate the coexisting densities with enough accuracy . on approaching the critical region
this problem becomes even worst . in order to better investigate the coexistence region and in particular the possible existence of a liquid - liquid phase coexistence we carried out extensive mes calculations . from our isotherms calculations
we have roughly estimated the region where the onset of the bimodal shape of the particles number distribution p(n ) can be observed . in our analysis
we started from the double - peaked p(n ) obtained at the temperature @xmath38 .
the histogram reweighting technique allowed us to estimate the coexistence liquid and vapor densities and chemical potential at this temperature .
then by making use of the mes procedure we investigated a wide range of subcritical temperatures . in fig .
[ fig : istog ] we report the distributions p(n ) for the thermodynamical states investigated starting from @xmath38 .
[ fig : istog ] at higher temperatures the liquid and gas have close densities and the double - peaked @xmath25 distribution shows two partially overlapped peaks . as the temperature decreases the difference between the coexisting vapor and liquid densities becomes more marked . in the intermediate temperature regime
we do not observe the occurrence of a third peak associated with a second coexisting liquid phase .
we will comment more extensively about this point in the following .
table [ tab:1 ] summarizes the temperatures at which we performed our simulations together with the values of the coexistence densities .
the mes algorithm allowed us to follow the system down to @xmath39 , where the liquid density is more than 10 times higher than the vapor one .
.the peak densities corresponding to the number particles distributions shown in fig .
[ fig : istog ] .
all quantities are in lennard - jones units . [ cols="^,^,^,^",options="header " , ] snapshots from our simulations for the coexistence liquid and vapor configurations at one temperature are shown in fig .
[ fig : gas ] and fig .
[ fig : liq ] . in both the coexisting phases the spatial distribution of the fluid molecules is inhomogeneous and the pores of the aerogel are far from being uniformly filled by the fluid .
[ fig : gas ] [ fig : liq ] fig .
[ fig : coe ] shows the temperature versus density phase diagram we obtained from the peaks locations in the p(n ) distributions depicted in fig .
[ fig : istog ] , compared with that of the bulk fluid @xcite .
we find that the confined fluid phase diagram is substantially modified by the presence of the aerogel : both the critical temperature and density are lower than in the bulk and the range of the vapor - liquid coexistence curve is much less extended .
we recall that in the bulk the critical parameters are : @xmath46 , @xmath47 and @xmath48 @xcite .
our findings about the gas - liquid coexistence properties are in qualitative agreement with previous computer simulation studies on lennard - jones fluids confined in random spheres matrices with purely repulsive adsorbent - adsorbate interactions . @xcite .
[ fig : coe ] none - the - less the phase diagram reported in fig .
[ fig : coe ] does not present a liquid - liquid coexistence region .
it shows instead a well defined shoulder on the liquid side boundary for intermediate temperatures in the range investigated . we must however stress that confining primary particles in real aerogels are substantially larger than the fluid molecules , therefore a study on a much larger system
could help to shed light on this issue . in order to better define the nature of the liquid phase for the thermodynamic points corresponding to this shouldering , in fig [ fig : liquid ]
we show a blow up of the liquid peaks of the p(n ) distributions centered around the intermediate temperature region . [
fig : liquid ] the liquid peaks generating the shoulder in the coexistence curve are indicated with bold solid lines .
the liquid peaks shape shows a peculiar behavior with varying temperature . for temperatures ranging from near the critical point to about @xmath49 ,
the shape of the liquid peaks is symmetric around the peaks maximum as observed for the bulk @xcite . for the reduced temperature corresponding to the onset of the shoulder , @xmath50 ( the first bold curve on the left )
, the liquid peak shows a bump on its right side that renders its shape slightly asymmetric .
we observe that the position of this bump roughly corresponds to the liquid peak maximum of the next @xmath25 distribution , corresponding to @xmath51 . for @xmath51 ( the second bold curve starting from the left )
it is conversely observed a bump located on the left side of the liquid peak at nearly the same position of the peak maximum measured at @xmath50 .
we additionally note that the main peaks of these two curves show a more pronounced separation from each other compared with the other curves , in spite of being the temperature jump between one curve and the next approximatively similar for all the curves investigated .
the next two liquid peaks , corresponding to the end of the shouldering , @xmath52 and @xmath53 , are less asymmetric and closer than the two previous ones . for temperatures lower than @xmath53 , after the shoulder in the phase diagram , the shape of the liquids peaks returns to be symmetric and the liquid densities increase regularly on lowering temperature , similar to the high temperatures range . from these observations we can infer that at high temperatures the gas phase coexists with a medium density liquid . correspondingly the coexistence curve decreases regularly with temperature . in the intermediate temperature range
the liquid phase starts showing high density fluctuations and appears to coexists with a slightly more dense liquid .
this region is characterized by an asymmetry in the liquid peaks of the @xmath25 and correspondingly a shouldering of the phase diagram , that mark a crossover to a slightly more dense phase . for lower temperatures the higher density
liquid becomes the thermodynamically favoured phase coexisting with the vapor and the coexistence curve resumes a regular descent .
finally we report the liquid - vapor coexistence curve in the ( @xmath18,@xmath19 ) plane as obtained from our multicanonical simulations . according to the equal peak weight criterion , described in sec . 3
, the location of the liquid - vapor saturation line in this plane can be obtained looking for the values of temperature and chemical potential at which the bimodal distributions @xmath25 have equal area under the two peaks .
[ fig : mt ] fig .
[ fig : mt ] shows the temperature behavior of the coexistence chemical potentials at which we obtained the p(n ) distributions reported in fig .
[ fig : istog ] .
analogous to the bulk we observe a monotonic regular behavior .
similar to the temperature versus density phase diagram , there is a substantial shrinkage of the range of chemical potentials at which we observe the coexistence of the liquid and gas phases with respect to the bulk .
in fact in the bulk @xcite for a similar interval of temperatures as ours , i.e. @xmath54 , a range @xmath55 is found , while in our confined system @xmath56 , more than six times lower .
we have performed a computer simulation study of the phase behavior of a lennard - jones fluid adsorbed in a highly porous fractal aerogel .
the gel environment has been generated with the dlca algorithm @xcite in order to obtain a more realistic confining structure than those reported in literature . besides the confined fluid phase diagram has been calculated by performing multicanonical ensemble simulations in the framework of the monte carlo grand canonical ensemble technique @xcite .
the proper bias has been found by applying the single - histogram reweighting technique @xcite .
this procedure is particularly suitable for analysing the transition state of the first order phase transitions in great detail .
we found that the phase diagram of the confined fluid is substantially modified with respect to the bulk by the presence of the gel : the critical temperature and density are lower and the liquid - gas coexistence curve is much narrower than in the bulk .
our findings about the vapor - liquid coexistence properties are in qualitative agreement with experimental observations and previous theoretical studies .
we have not found a clear evidence for a second liquid - liquid coexistence region in the fluid phase diagram .
the presence of a liquid - liquid phase coexistence has been sometimes reported , depending on the particular configuration generated for the adsorbent , random matrices of spheres @xcite .
our coexistence curve shows a shoulder on the liquid side boundary in the intermediate temperature range .
for the same temperatures the liquid peaks shape of our particles number distributions are highly asymmetric and show a bump . however , since we do not find two distinctly resolved peaks , we must conclude that in our case coexistence between two liquid phases is never observed .
we do observe instead a crossing of the liquid from a lower to an higher density phase upon decreasing temperature .
we have employed in this study a very sophisticated technique for an accurate location of the coexistence curve . due to the fractal structure of the confining environment
the results should in principle not depend on the realization .
however we can not exclude that finite size effects could be present . only extensive analysis with increasing sizes
can confirm how sensitive the results are to the matrix realization . in order to gain deep insight about the modification of the phase diagram of fluids confined in aerogels the modeling of the adsorbent environment
can be further improved to make contact with the experimental situations .
it would be interesting to explore in a more carefully designed confining structure with a network of multishaped voids if the two liquid phases will become distinguishable .
in particular the liquid - liquid coexistence is smeared out in our present simulation , where we observe only a shouldering in the phase diagram .
it would be valuable to perform simulations on even larger systems to better investigate the role of the fractal behavior of the confining medium and perform an extensive analysis of finite size effects .
, 1573 ( 1999 ) . ,
60 ( 1984 ) . , 2567 ( 1990 ) . , 2754 ( 1991 ) . , 632 ( 1992 )
, 954 ( 1993 ) . ,
5922 ( 1995 ) . , 1865 ( 1987 ) . ,
1954 ( 1988 ) . , 804 ( 1989 ) . ,
1185 ( 1991 ) . , 3535 ( 1992 ) .
, 2207 ( 1994 ) . f. brochard and p. g.
de gennes , _ j. phys.(france ) lett .
_ * 44 * , 785 ( 1983 ) ; p.g .
de gennes , _ j. phys
. chem . _ * 88 * , 6469 ( 1984 ) .
, 537 ( 1988 ) ; w. g. madden _ j. chem . phys .
_ * 96 * , 5422 ( 1992 ) . , 264 ( 1997 ) . , 11289 ( 2001 ) .
, 056124 ( 2002 ) . , r29 ( 1996 ) ; k. s. page and p. a. monson _ phys .
e _ * 54 * , 6557 ( 1996 ) . , 5495 ( 1999 ) .
, 7231 ( 2000 ) .
d. w. schaefer and k. d. keefer , _ phys .
lett . _ * 56 * , 2199 ( 1986 ) . , 269 ( 1986 ) . , 9345 ( 1993 ) . ,
6006 ( 1994 ) . ,
r1305 ( 1994 ) . , l435 ( 1985 ) . ,
9 ( 1992 ) . , 602 ( 1995 ) . , 1147 ( 2001 ) .
oxford university press ( 1987 ) . academic press , boston ( 1996 ) .
, 2635 ( 1988 ) ; a. m. ferrenberg and r. h. swendsen _ phys
_ * 63 * , 1195 ( 1989 ) ; a. m. ferrenberg and r. h. swendsen _ computers in physics _ * 3*(5 ) , 101 ( 1989 ) . ,
055701 ( 2001 ) . | we present a grand canonical monte carlo simulation study of the phase diagram of a lennard - jones fluid adsorbed in a fractal and highly porous aerogel .
the gel environment is generated from an off - lattice diffusion limited cluster - cluster aggregation process .
simulations have been performed with the multicanonical ensemble sampling technique .
the biased sampling function has been obtained by histogram reweighting calculations . comparing the confined and the bulk system liquid - vapor coexistence curves we observe a decrease of both the critical temperature and density in qualitative agreement with experiments and other monte carlo studies on lennard - jones fluids confined in random matrices of spheres . at variance with these numerical studies
we do not observe upon confinement a peak on the liquid side of the coexistence curve associated with a liquid - liquid phase coexistence . in our case
only a shouldering of the coexistence curve appears upon confinement .
this shoulder can be associated with high density fluctuations in the liquid phase .
the coexisting vapor and liquid phases in our system show a high degree of spatial disorder and inhomogeneity . |
that hadrons can change their properties and couplings in the nuclear medium has been well known to nuclear physicists since the days of the delta - hole model that dealt with the changes of the properties of the pion and delta - resonance inside nuclei @xcite . due to the predominant @xmath6-wave interaction of pions with nucleons
one observes here a lowering of the pion branch for small but finite pion momenta , which increases with the nucleon - density .
more recently , experiments at the fsr at gsi have shown that also the pion rest mass in medium differs from its value in vacuum @xcite .
this is interesting since there are also recent experiments @xcite that look for the in - medium changes of the @xmath7 meson , the chiral partner of the pion .
any comparison of scalar and pseudoscalar strength could thus give information about the degree of chiral symmetry restoration in nuclear matter .
in addition , experiments for charged kaon production at gsi @xcite have given some evidence for the theoretically predicted lowering of the @xmath8 mass in medium and the ( weaker ) rising of the @xmath9 mass .
state - of - the - art calculations of the in - medium properties of kaons have shown that the usual quasi - particle approximation for these particles is no longer justified inside nuclear matter where they acquire a broad spectral function @xcite . at higher energies , at the cern sps and most recently at the brookhaven rhic , in - medium changes of vector mesons
have found increased interest , mainly because these mesons couple strongly to the photon so that electromagnetic signals could yield information about properties of hadrons deeply embedded into nuclear matter .
indeed , the ceres experiment @xcite has found a considerable excess of dileptons in an invariant mass range from @xmath10 mev to @xmath11 mev as compared to expectations based on the assumption of freely radiating mesons .
this result has found an explanation in terms of a shift of the @xmath12 meson spectral function down to lower masses , as expected from theory ( see , e.g. , @xcite ) .
however , the actual reason for the observed dilepton excess is far from clear .
both models that just shift the pole mass of the vector meson as well as those that also modify the spectral shape have successfully explained the data @xcite ; in addition , even a calculation that just used the free radiation rates with their often quite large
experimental uncertainties was compatible with the observations @xcite .
there are also calculations that attribute the observed effect to radiation from a quark - gluon plasma @xcite . while all these quite different model calculations tend to explain the data , though often with some model assumptions , their theoretical input is sufficiently different as to make the inverse conclusion that the data prove one or another of these scenarios impossible .
one of the authors has , therefore , already some years ago proposed to look for the theoretically predicted changes of vector meson properties inside the nuclear medium in reactions on normal nuclei with more microscopic probes @xcite . of course
, the nuclear density felt by the vector mesons in such experiments lies much below the equilibrium density of nuclear matter , @xmath13 , so that naively any density - dependent effects are expected to be much smaller than in heavy - ion reactions . on the other hand , there is a big advantage to these experiments : they proceed with the spectator matter being close to its equilibrium state .
this is essential because all theoretical predictions of in - medium properties of hadrons are based on an model in which the hadron ( vector meson ) under investigation is embedded in cold nuclear matter in equilibrium and with infinite extension .
however , a relativistic heavy - ion reaction proceeds at least initially far from equilibrium .
even if equilibrium is reached in a heavy - ion collision this state changes by cooling through expansion and particle emission and any observed signal is built up by integrating over the emissions from all these different stages of the reaction .
another in - medium effect arises when particles are produced by high - energy projectiles inside a nuclear medium .
a major experimental effort at rhic experiments has gone into the observation of jets in ultra - relativistic heavy - ion collisions and the determination of their interaction with the surrounding quark or hadronic matter @xcite .
a complementary process is given by the latest hermes results at hera for high - energy electroproduction of hadrons off nuclei @xcite .
again , the advantage of the latter experiment is that the nuclear matter with which the interactions happen is at rest and in equilibrium . in this lecture note
we summarize results that we have obtained in studies of observable consequences of in - medium changes of hadronic spectral functions as well as hadron formation in reactions of elementary probes with nuclei .
we demonstrate that the expected in - medium sensitivity in such reactions is as high as that in relativistic heavy - ion collisions and that in particular photonuclear reactions present an independent , cleaner testing ground for assumptions made in analyzing heavy - ion reactions .
a large part of the current interest in in - medium properties of hadrons comes from the hope to learn something about quarks in nuclei .
indeed , a very simple estimate shows that the chiral condensate in the nuclear medium is in lowest order in density given by @xcite @xmath14 here @xmath15 is the _ scalar _ density of the hadron @xmath16 in the nuclear system and @xmath17 the so - called sigma - commutator that contains information on the chiral properties of @xmath16 .
the sum runs over all hadronic states .
while ( [ qbarq ] ) is nearly exact , its actual applicability is limited because neither the sigma - commutators of the higher lying hadrons nor their scalar densities are known . however , at very low temperatures close to the groundstate these are accessible . here
@xmath18 so that the condensate drops linearly with the nuclear ( vector ) density .
this drop can be understood in physical terms : with increasing density the hadrons with a chirally symmetric phase in their interior fill in more and more space in the vacuum with its spontaneously broken chiral symmetry .
note that this is a pure volume effect ; it is there already for a free , non - interacting hadron gas .
how this drop of the scalar condensate translates into observable hadron masses is not uniquely prescribed .
the only rigorous connection is given by the qcd sum rules that relates an integral over the hadronic spectral function to a sum over combinations of quark- and gluon - condensates with powers of @xmath19 .
the starting point for this connection is the current - current correlator @xmath20 where the current is expressed in terms of quark - field operators .
this correlator can be decomposed @xcite @xmath21 the qcd sum rule then reads @xmath22 with @xmath23 and some subtraction constants @xmath24 . here
@xmath25 represents a wilson s operator expansion of the current - current correlator in terms of quark and gluon degrees of freedom in the space - like region . on the other hand , @xmath26 in ( [ eq : opehadr ] )
is the same object for time - like momenta , represented by a parametrization in terms of hadronic variables .
the dispersion integral connects time- and space - like momenta .
( [ eq : opehadr ] ) also connects the hadronic with the quark world .
it allows after the technical step of a borel transformation to determine parameters in a hadronic parametrization of @xmath26 by comparing the lhs of this equation with its rhs . the operator product expansion of @xmath25 on the lhs involves quark- and gluon condensates @xcite ; of these only the two - quark condensates
are reasonably well known , whereas our knowledge about already the four - quark condensates is rather sketchy .
leupold et al .
@xcite have shown that using a realistic parametrization of the hadronic spectral function on the rhs of ( [ eq : opehadr ] ) @xmath27 where @xmath28 denotes the threshold between the low energy region described by a spectral function for the lowest lying resonance , @xmath29 , and the high energy region described by a continuum calculated from perturbative qcd .
the second term on the rhs of ( [ eq : pihadans ] ) represents the qcd perturbative result that survives when @xmath30 . using the measured , known vacuum spectral function for @xmath31 allows one to obtain information about the condensates appearing on the lhs of ( [ eq : opehadr ] ) . on the other hand , modelling the density - dependence of the condensates yields information on the change of the hadronic spectral function when the hadron is embedded in nuclear matter .
since the spectral function appears under an integral the information obtained is not very precise .
however , leupold et al . have shown @xcite that the qcdsr provides important constraints for the hadronic spectral functions in medium , but it does not fix them .
recently kaempfer et al have turned this argument around by pointing out that measuring an in - medium spectral function of the @xmath4 meson could help to determine the density dependence of the higher - order condensates @xcite .
thus models are needed for the hadronic interactions .
the quantitatively reliable ones can at present be based only on classical hadrons and their interactions .
indeed , in lowest order in the density the mass and width of an interacting hadron in nuclear matter at zero temperature and vector density @xmath32 are given by ( for a meson , for example ) @xmath33 here @xmath34 is the forward scattering amplitude for a meson with energy @xmath35 on a nucleon .
the width @xmath36 denotes the free decay width of the particle . for the imaginary part this is nothing other than the classical relation @xmath37 for the collision width , where @xmath7 is the total cross section
this can easily be seen by using the optical theorem .
note that such a picture also encompasses the change of the chiral condensate in ( [ qbarq ] ) , obtained there for non - interacting hadrons . if the spectral function of a non - interacting hadron changes as a function of density , then in a classical hadronic theory , which works with fixed ( free ) hadron masses , this change will show up as an energy - dependent interaction and is thus contained in any empirical phenomenological cross section .
unfortunately it is not a - priori known up to which densities the low - density expansion ( [ trho ] ) is useful .
post et al .
have recently investigated this question in a coupled - channel calculation of selfenergies @xcite .
their analysis comprises pions , @xmath0-mesons and @xmath12-mesons as well as all baryon resonances with a sizeable coupling to any of these mesons .
[ inmedself ] shows a typical selfenergy graph in their calculation .
of particular importance are the @xmath38 , the @xmath1 and the @xmath39 resonances for the pion , the @xmath0-meson and the @xmath12-meson , respectively .
the authors of @xcite find that already for densities less than @xmath40 the linear scaling of the selfenergies inherent in ( [ trho ] ) is badly violated for the @xmath12 and the @xmath41 mesons , whereas it is a reasonable approximation for the @xmath0 meson , but even in the latter case already at @xmath40 the selfenergy does not scale exactly with @xmath12 .
reasons for this deviation from linearity are fermi - motion , pauli - blocking , selfconsistency and short - range correlations . for different mesons different sources of the discrepancy prevail : for the @xmath12 and @xmath0 mesons the iterations act against the low - density theorem by inducing a strong broadening for the @xmath39 and a slightly repulsive mass shift for the @xmath1 nucleon resonances to which the @xmath12 and the @xmath0 meson , respectively , couple .
the effects of self - consistency on the spectral function of the @xmath12-meson can nicely be seen in fig .
[ rhospectral ] .
the figure shows that the spectral function of a @xmath12 meson at rest in the nuclear medium ( upper left , solid curve ) exhibits a remarkable structure that is generated by the strong coupling to the @xmath39 resonance . while this structure disappears at the highest momentum considered here ,
the spectral function for the transverse vector meson ( lower left ) is still considerably different from that of a @xmath12 in vacuum .
notice also that the longitudinal spectral function ( right column ) retains more of the original vector meson structure .
traditionally , heavy - ion reactions are assumed to be the reactions of choice for an investigation of in - medium properties .
after all , here very high nuclear densities ( @xmath42 ) can be reached with present day s accelerators so that any density - dependent effect would be magnified .
however , any explanations of observed effects always suffer from an inherent inconsistency : the observed signal , on one hand , necessarily integrates over many different stages of the collision nonequilibrium and equilibrium , the latter at various densities and temperatures . on the other hand ,
the theoretical input is always calculated under the simplifying assumption of a hadron in nuclear matter in equilibrium .
we have therefore looked for possible effects in reactions that proceed much closer to equilibrium , i.e. reactions of elementary probes such as protons , pions , and photons on nuclei .
of course , the densities probed in such reactions are always @xmath43 , with most of the nucleons actually being at about @xmath40 . on the other hand ,
the target is obviously stationary and the reaction proceeds much closer to ( cold ) equilibrium than in a relativistic heavy - ion collision .
it is thus a quantitative question if any observable effects of in - medium changes of hadronic properties survive if the densities probed are always @xmath43 . with
the aim of answering this question we have over the last few years undertaken a number of calculations for proton- @xcite , pion- @xcite and photon- @xcite induced reactions . all of them have one feature in common : they treat the final state incoherently in a coupled channel transport calculation that allows for elastic and inelastic scattering of , particle production by and absorption of the produced hadrons ( vector mesons ) .
a new feature of these calculations is that vector mesons with their correct spectral functions can actually be produced and transported consistently .
this is quite an advantage over earlier treatments @xcite in which the mesons were always produced and transported with their pole mass and their spectral function was later on folded in only for their decay . in all photonuclear calculations above energies of about 1 gev
we account for coherent initial state interactions of the incoming photon which lead to shadowing . in an interaction with a bound nucleon
the shadowed incoming photon then produces a hadronic final state which cascades through the nucleus @xcite .
the incoherent final state interactions are treated as just described by means of a semi - classical coupled - channel transport theory . for details of the model
see ref . @xcite and references therein .
we first look at the prospects of using reactions with hadronic final states and discuss the photo - production of @xmath0-mesons on nuclei as a first example .
these mesons are unique in that they are sensitive to one dominating resonance only ( the @xmath1 ) so that one may hope to learn something about the properties of this resonance inside nuclei .
experiments for this reaction were performed both by the taps collaboration @xcite and for higher energies at kek @xcite .
estimates of the collisional broadening of the @xmath1 resonance have given a collision width of about 35 mev at @xmath13 @xcite .
the more recent , and more refined , selfconsistent calculations of @xcite give a very similar value for this resonance .
in addition , a dispersive calculation of the real part of the selfenergy for the resonance at rest gives only an insignificant shift of the resonance position .
thus any momentum - dependence of the selfenergy as observed in photon - nucleus data can directly be attributed to binding energy effects @xcite .
the results of this calculation @xcite are shown in fig .
[ etaphot ] . the dashed line in fig .
[ etaphot ] gives the results of a calculation employing a momentum - independent potential for the @xmath1 resonance with the same depth at @xmath13 as for the nucleons .
it is clearly seen that it rises too steeply and peaks at a two low energy compared to the data .
the solid line , that describes all available data over a wide energy range very well , employs a momentum - dependence in the nucleon and resonance potentials taken from earlier fits to heavy - ion reaction data . for the present discussion
the relevant feature of this potential is that it vanishes for momenta around 800 mev .
thus , for photon energies around this values , close to the peak position in the data , the resonance is no longer bound whereas the hit nucleon is bound by about 75 mev at @xmath13 .
thus , effectively the resonance is shifted upwards by this amount ; this is just what the data show .
in addition , fig .
[ etaphot ] also contains a dotted line , exhibiting the effects of the in - medium width of the @xmath1 resonance . since the latter is small , also its effect on the eta - production cross section is small . as a second example
we look into the photoproduction of @xmath2-mesons in nuclei .
our calculations of this reaction have been motivated by the idea to learn about the @xmath2 in - medium properties by measuring the @xmath3 mass distribution @xcite .
it has been proposed to perform such an experiment at the spring8 facility in osaka @xcite . in principle
the @xmath2 meson forms a unique probe since it is the only resonance decaying into @xmath3 pairs in the considered mass region .
theoretical approaches to the @xmath2 in - medium properties find a broadening of the @xmath2 meson of up to 40 mev at nuclear matter density and vanishing @xmath2 momentum @xcite . while this value exceeds the vacuum value of 4.4 mev by about one order of magnitude , a possible mass shift of the @xmath2
is expected to be of the order of only @xmath44 mev to @xmath45 mev @xcite so that the @xmath2 meson is expected to survive as a proper quasiparticle even at nuclear matter density @xmath13 @xcite .
thus the expected modification might lead to observable consequences because of the small width of the @xmath2 meson .
only vector mesons decaying inside the nucleus may reveal informations about their in - medium properties .
hence , the decay length of these vector mesons @xmath46 has to be of the order of the nuclear radius . in order to produce @xmath2 mesons with the required low momenta in the target rest frame
the incident photon energy has to be chosen not too close above threshold ( @xmath47 1.5 gev ) .
simulations at higher photon energies require to include also inclusive photoproduction processes .
this is illustrated in fig .
[ phi01 ] , which shows the momentum distribution of @xmath2 mesons with ( solid line ) and without fsi ( dashed ) from a nuclear target ( @xmath48ca ) .
it is clearly seen that the cross section at small @xmath2 momenta is dominated by the inclusive production processes . in the simulations inclusive
photoproduction of vector mesons has been accounted for by means of the hadronic event generator fritiof employing the vector meson dominance model in order to describe the photon - hadron coupling @xcite .
also visible from fig .
[ phi01 ] is the fact , that the fsi of the @xmath2 , through their non - absorptive part , produce an accumulation of @xmath2 mesons in the low - momentum part of the spectrum .
the number of @xmath2 mesons decaying in the medium is greatly enhanced by this effect . including collisional
broadening by means of the low - density expansion ( [ trho ] ) and applying a rather severe momentum cutoff to the three momentum of the detected @xmath3 pair , we find an observable broadening of the invariant mass spectrum by almost a factor of two , see fig .
[ phi02 ] . however , accounting also for the electromagnetic potential which already in @xmath48ca has a depth of 11 mev , the average decay density of reconstructed @xmath2 mesons decreases considerably . as a consequence
the obtained invariant mass spectrum again is very close to the free one ( dotted in fig .
[ phi02 ] ) . also the dropping @xmath2 mass ( 30 mev at @xmath13 in the simulation ) has no remarkable effect , since as long as the @xmath9 and @xmath8 properties are not changed in the medium @xmath2 mesons with considerably reduced masses do not decay into @xmath3 pairs .
this also leads to a reduced average decay density .
as already mentioned , also the kaon properties receive a rather dramatic change in the nuclear environment @xcite .
kaons from @xmath2 decays that propagate through a potential gradient change their momenta and , hence , also the invariant mass of any @xmath3 pair .
this effect leads to an accumulation of strength on the high mass side of the invariant mass spectrum , see fig .
[ phi03 ] . in summary , we find , that due to the medium modification of the @xmath2 decay products ( the kaons ) one can not learn anything about the @xmath2 properties in the @xmath49 reaction , but rather about the existence of an in - medium change of the @xmath9 and @xmath8 mesons .
in contrast to the results for the @xmath2 meson , the situation looks more promising for the @xmath4 meson @xcite . on the one hand ,
the @xmath4 is much lighter than the @xmath2 and can , therefore , be produced with lower momentum inside nuclei .
on the other hand , the @xmath5 decay mode is not influenced by the electromagnetic potential and also fsi effects might be less pronounced because of the weak photon - nucleon coupling .
an experiment measuring the @xmath50 reaction is presently being analyzed by the taps / crystal barrel collaborations at elsa @xcite .
the varying theoretical predictions for the @xmath4 mass ( 640 - 765 mev ) @xcite and width ( up to 50 mev ) @xcite in nuclear matter at rest encourage the use of such an exclusive probe to learn about the @xmath4 spectral distribution in nuclei .
simulations have been performed at 1.2 gev and 2.5 gev photon energy , which cover the accessible energies of the taps / crystal barrel experiment .
after reducing the combinatorial and rescattering background by applying kinematic cuts on the outgoing particles , we obtain rather clear signals for a dropping @xmath4 mass inside nuclei . the solid line in fig .
[ omega01 ] shows results of a calculation including both collisional broadening and a lowering of the @xmath4 mass .
it exhibits a shoulder on the low - mass side of the peak and a clear excess there over the proton distribution , also shown in the figure .
this has to be compared to the dashed line , which includes collisional broadening only .
therefore , in this particular case it is possible to disentangle rather trivial in - medium changes ( collisional broadening ) from more interesting medium - modifications ( dropping mass ) , which may be related to a change of the chiral condensate at finite nuclear density @xcite . employing the low - density expansion of eq .
( [ trho ] ) to determine the effective in - medium mass of the @xmath4 one obtains a momentum dependence which is driven by the energy dependence of the @xmath4-nucleon total cross section .
such a simple - minded approach yields a considerably reduced mass for @xmath4 mesons at rest and a fast increase of the effective mass for @xmath4 mesons moving with finite velocity , arriving rather fast at values close to the @xmath4 vacuum mass @xcite .
the results of simulations adopting this momentum dependent mass are also shown in fig .
[ omega01 ] ( dash - dotted line ) .
as expected the results of these calculations are again very similar to the results obtained without any real @xmath4 potential .
measurements with varying momentum cuts may help to disentangle the density and momentum dependence of the in - medium self - energy of the @xmath4 meson .
dileptons , i.e. electron - positron pairs , in the outgoing channel are an ideal probe for in - medium properties of hadrons since they in contrast to hadronic probes experience no strong final state interaction . a first experiment to look for these dileptons in heavy - ion reactions was the dls experiment at the bevalac in berkeley @xcite . later on , and in a higher energy regime , the ceres experiment has received a lot of attention for its observation of an excess of dileptons with invariant masses below those of the lightest vector mesons @xcite .
explanations of this excess have focused on a change of in - medium properties of these vector mesons in dense nuclear matter ( see e.g.@xcite ) .
the radiating sources can be nicely seen in fig .
[ ceres ] that shows the dilepton spectrum obtained in a low - energy run at 40 agev together with the elementary sources of dilepton radiation .
the figure exhibits clearly the rather strong contributions of the vector mesons both direct and through their dalitz decay at invariant masses above about 500 mev .
if this strength is shifted downward , caused by an in - medium change of the vector - meson spectral functions , then the observed excess can be explained as has been shown by various authors ( see e.g. @xcite for a review of such calculations ) .
the results of a new calculation , based on the spectral function in fig .
[ rhospectral ] , are shown in fig .
[ tdilept ] .
the strong amplification of the dilepton rate at small invariant masses is caused by the photon propagator , which contributes @xmath51 to the cross section .
any strength in the spectral function at finite density at small @xmath52 is thus dramatically amplified . in view of the uncertainties in interpreting these results discussed earlier
we have studied the dilepton production in reactions on nuclear targets involving more elementary projectiles .
it is not _ a priori _ hopeless to look for in - medium effects on dilepton production in ordinary nuclei : even in relativistic heavy - ion reactions that reach baryonic densities of the order of 3 - 10 @xmath13 many observed dileptons actually stem from densities that are much lower than these high peak densities .
transport simulations have shown @xcite that even at the ceres energies about 1/2 of all dileptons come from densities lower than @xmath53 .
this is so because in such reactions the pion - density gets quite large in particular in the late stages of the collision , where the baryonic matter expands and its density becomes low again .
correspondingly many vector mesons are formed ( through @xmath54 ) late in the collision and their decay to dileptons thus happens at low baryon densities .
a typical result of such a calculation for the dilepton yield after removing the bethe - heitler component is given in fig .
[ fige+e- ] . comparing this figure with fig .
[ ceres ] shows that in a photon - induced reaction at 1 - 2 gev photon energy exactly the same sources , and none less , contribute to the dilepton yield as in relativistic heavy - ion collisions at 40 agev !
the question now remains if we can expect any observable effect of possible in - medium changes of the vector meson spectral functions in medium in such an experiment on the nucleus where due to surface effects the average nucleon density is below @xmath13 .
this question is answered , for example , by the results of fig.[figdlim ] .
this figure shows the dilepton spectra to be expected if a suitable cut on the dilepton momenta is imposed ; with this cut slow vector mesons are enriched . in the realistic case shown on the right , which contains both a collision broadening and a mass shift , it is obvious that a major signal is to be expected : in the heavy nucleus @xmath55 the @xmath4-peak has completely disappeared from the spectrum .
the sensitivity of such reactions is thus as large as that observed in ultrarelativistic heavy - ion reactions .
an experimental verification of this prediction would be a major step forward in our understanding of in - medium changes .
it would obviously present a purely hadronic base - line to all data on top of which all fancier explanations of the ceres effect in terms of radiation from a qgp and the such would have to live .
high energy photo- and electroproduction of hadrons on complex nuclei offers a promising tool to study the physics of hadron formation .
the interaction of the ( virtual ) photon with a nucleon leads to the production of several high energy particles . due to the soft momentum scales involved in the hadronization process
the formation of the final hadrons is of nonperturbative nature and not well understood .
a simple estimate for the hadron formation time can be given by the time that the constituents need to travel a hadronic radius .
it should therefore be of the order of 0.50.8 fm / c in the rest frame of the hadron .
due to time dilatation the formation lengths of high energetic particles can reach distances in the lab frame that are of the order of the nuclear radius . thus nuclei can serve as a kind of microdetector for the determination of formation times as well as the interactions while the final hadron is still being formed . obviously , the latter is closely related to color transparency .
the hermes collaboration @xcite has investigated the multiplicities of different hadron species in deep inelastic scattering off various nuclei . here
the photon - energies are of the order of 1020 gev , with rather moderate @xmath56 gev@xmath57 . a similar experiment at somewhat smaller energies
is currently performed by the clas collaboration at jefferson lab @xcite .
after the planned energy upgrade they will reach photon energies of the order of 29 gev and similar virtualities as in the hermes experiment .
the interpretation of the observed multiplicity spectra reach from a possible rescaling of the quark fragmentation function at finite nuclear density combined with hadron absorption in the nuclear environment @xcite to purely partonic energy loss through induced gluon radiation of the struck quark propagating thorugh the nucleus @xcite .
note that a detailed understanding of energy loss and hadron attenuation in cold nuclear matter can provide a useful basis for the interpretation of jet quenching observed in ultra - relativistic heavy ion collisions at rhic .
in our approach @xcite we assume that the virtual photon interacts with the bound nucleon either directly or via one of its hadronic fluctuations .
the hadronic components of the photon are thereby shadowed according to the method developed in ref .
@xcite which allows for a clean - cut separation of the coherent initial state interactions of the photon and the incoherent final state interactions of the reaction products .
the photon nucleon interaction leads to the excitation of one or more strings which fragment into color - neutral prehadrons due to the creation of quark antiquark pairs from the vacuum . as discussed in a recent work by kopeliovich et al .
@xcite the production time of these prehadrons is very short .
for simplicity we therefore set the production time to zero in our numerical realization .
these prehadrons are then propagated using our semi - classical coupled - channel transport theory .
after a formation time , which we assume to be a constant @xmath58 in the restframe of the hadron , the hadronic wave function has build up and the reaction products behave like usual hadrons . during the formation time we distinguish between the so - called leading prehadrons which contain ( anti-)quarks from the struck nucleon or the hadronic component of the photon and the non - leading hadrons that solely contain quarks and antiquarks created from the vacuum .
the leading prehadrons can be considered as the debris of the initial hadronic state and each interacts with an effective cross section @xmath59 during the formation time .
the sum of the leading hadron cross sections should resemble approximately that of the initial hadronic state .
the nonleading prehadrons that are newly - created from the vacuum are assumed to be noninteracting during @xmath58 .
each time when a new hadron has formed the total effective cross section of the final state then rises like in the approach of ref .
@xcite . due to our coupled channel - treatment of the final state interactions the ( pre)hadrons
might not only be absorbed in the nuclear medium but can produce new particles in an interaction , thereby shifting strength from the high to the low energy part of the hadron spectrum .
our calculated multiplicity ratios of charged hadrons are shown in fig .
[ formtime ] in comparison with the experimental hermes data @xcite .
the figure clearly shows that the observed hadron multiplicities can be described only with formation times @xmath60 fm .
the curves obtained with larger formation times all lie very close together .
this is a consequence of the finite size of the target nucleus : if the formation time is larger than the time needed for the preformed hadron to transverse the nucleus , then the sensitivity to the formation time is lost . in @xcite
we have also shown that the @xmath61-dependence on the left side exhibits some sensitivity to the interactions of the leading hadrons during the formation ; the curves show in fig .
[ formtime ] are obtained with a down scaled leading hadron cross section of @xmath62 , where @xmath63 is the normal hadronic interaction cross section .
in this lecture note we have illustrated with a few examples that photonuclear reactions can yield information that is important and relevant for an understanding of high density high temperature phenomena in ultrarelativistic heavy - ion collisions .
special emphasis was put in this article not so much on the theoretical calculations of hadronic in - medium properties under simplified conditions , but more on the final , observable effects of any such properties .
we have discussed both reactions with hadronic final states as well as those with fsi - free lepton states .
while in the former reactions the fsi on the final decay products usually is strong enough to mask the in - medium spectral properties of the hadron under study , in the latter case ( semi - hadronic or leptonic final states ) the situation is much more encouraging .
we have , for example , shown that in photonuclear reactions in the 1 - 2 gev range the expected sensitivity of dilepton spectra to changes of the @xmath12- and @xmath4 meson properties in medium is as large as that in ultrarelativistic heavy - ion collisions .
we have also illustrated that the analysis of hadron production spectra in high - energy electroproduction experiments at hermes gives information about the interaction of forming hadrons with the surrounding hadronic matter .
this is important for any analysis that tries to obtain signals for a qgp by analysing high - energy jet formation in ultrarelativistic heavy - ion reactions .
this work has been supported by the deutsche forschungsgemeinschaft , the bmbf and gsi darmstadt .
p. muehlich et al . , , 024605 ( 2003 ) .
t. nakano et al . , ,
71 ( 2001 ) .
e. oset , a. ramos , , 527 ( 1997 ) .
t. hatsuda , s. h. lee , , 34 ( 1992 ) .
f. klingl , t. waas , w. weise , , 254 ( 1998 ) . j. schaffner , j. bondorf , i. n. mishustin , , 325 ( 1997 ) .
p. muehlich et al . , nucl - th/0310067 . j. g. messchendorp et al . , _ eur .
j. _ * a11 * , 95 ( 2001 ) .
f. klingl , t. waas , w. weise , , 299 ( 1999 ) .
b. friman , acta phys . polon . *
b29 * , 3195 ( 1998 ) .
t. weidmann et al . , , 919 ( 1999 ) ( 1999 ) g. e. brown , m. rho , , 2720 ( 1991 ) . | we first briefly review the theoretical basis for calculations of changes of hadronic properties in dense nuclear matter .
these changes have usually been investigated by means of relativistic heavy - ion reactions . here
we discuss that observable consequences of such changes can also be seen in more elementary reactions on nuclei .
particular emphasis is put on a discussion of actual observables in photonuclear reactions ; we discuss in detail @xmath0- and vector - meson production .
we show that photoproduction of @xmath0 s can yield essential information on in - medium properties of the @xmath1 resonance while the @xmath2 meson properties will probably not be accessible through the @xmath3 decay channel .
however , for @xmath4 mesons the @xmath5 decay channel , due to its reduced final state interaction , looks more promising in this respect .
completely free of final state interactions is dilepton production in the few gev range .
we show that the sensitivity of this decay channel to changes of hadronic properties in medium in photonuclear reactions on nuclei is as large as in ultrarelativistic heavy ion collisions .
finally we discuss that hadron production in nuclei at 10 20 gev photon energies can give important information on the hadronization process . |
photons of energies around 2 mev have a high probability to interact in germanium through compton scattering .
the mean free path of the process is a couple of centimeters . if a photon compton scatters only once inside a germanium detector , the recoiling electron deposits its energy most likely within a 1 mm range , resulting in a so - called single - site event ( sse ) .
if , in contrast , a photon interacts through pair production or scatters multiple times , energy can be deposited at different locations separated by typically a few centimeters , resulting in a so - called multi - site event ( mse ) .
the charge carriers created by the energy deposition in the germanium detector drift towards the anode and cathode of the detector . while the charge amplitude of the induced pulse is determined by the number of carriers ( thus by the energy deposited ) , the pulse shape is determined by the location(s ) of the energy deposition(s ) and thus the charge drifting times .
mses are expected to have more involved pulse shapes than sses , and thus , pulse shape analysis ( psa ) can be used to separate the two classes of events @xcite .
one application of psa is the background rejection in experiments searching for neutrinoless double - beta decay ( 0@xmath2@xmath3 ) in @xmath1ge - enriched detectors , such as the gerda experiment @xcite .
the expected 0@xmath2@xmath3 signal events have two electrons in the final state with a total energy of 2.039 mev . these are mostly sses .
a large fraction of the expected background events are induced by external photons with energy depositions around the q - value .
these events are expected to be predominantly mses which can be rejected by psa . in order to study and improve the performance of psa , sse- and mse - dominant data samples have to be collected independently of the pulse shape . in this paper
a method to collect single compton scattering events ( scs ) as an sse - dominant sample is investigated in more detail .
the energy of the scattered photon in an scs event can be calculated given the incoming photon energy and the scattering angle .
therefore , scs events can be collected by positioning a second germanium detector at a specific angle with respect to the first detector and using it to tag escaped photons with the correct energy @xcite .
if the incoming photon has an energy of 2.614 mev as emmitted by a @xmath5tl source , a photon compton scattered at 72@xmath6 has an energy of 575 kev .
this signature is used to tag the single recoiling electron inside the first germanium detector .
the energy in the event is equal to the germanium 0@xmath2@xmath3 q - value .
the location of the energy deposition of the electron within the detector volume is controlled by positioning the source and the second detector correspondingly .
another common method to collect an sse - dominant sample is to select the double - escape events ( dep ) @xcite .
the incoming photon interacts with the germanium detector through pair production and the two 511 kev photons from the positron annihilation escape the detector without further interaction .
the electron and positron mostly deposit their energies very locally and result in an sse .
another useful sample contains so - called single - escape events ( sep ) where only one 511 kev photon escapes .
the other photon mostly deposits its energy at locations different from those of the electron and positron .
thus , sep events provide an mse - dominant sample with energy deposition close to the 0@xmath2@xmath3 q - value . however , the dep events are not a perfect test sample for the expected 0@xmath2@xmath3 events . if the two photons escape the detector , the interaction point is more likely close to the detector surface as compared to scs events .
0@xmath2@xmath3 events , on the other hand , are distributed evenly within the detector volume .
in addition , dep and 0@xmath2@xmath3 events have different energies . a dep event induced by a 2.6 mev photon from a @xmath5tl source
has an energy of 1.59 mev , quite different from the 0@xmath2@xmath3 q - value . in these respects studies with scs samples
suffer less from systematic effects .
the experimental setup and the data collection are described in chapter [ chapter : experiment ] . the monte carlo simulation is also included in this chapter .
it is used to verify that the collected scs samples are sse - dominated . in chapter [ chapter : psa ] a psa package based on an artificial neural network ( ann ) is presented .
the training methods are described and the results given .
the experimental setup is illustrated in figure [ fig : setup ] . .
the dotted lines illustrate the segment boundaries . ]
the segmented germanium detector under study is a prototype detector for phase - ii of the gerda experiment @xcite .
the true coaxial 18-fold segmented @xmath7-type hpge dectector has a weight of 1.63 kg and the dimensions are 69.8 mm height and 75.0 mm diameter ; the inner hole has a diameter of 10.0 mm .
the segmentation scheme is 3-fold along the vertical axis and 6-fold in the azimuthal angle .
( see figure [ fig : setup ] ) .
signals from the 18 segments and the core of the detector are amplified by charge sensitive pre - amplifiers and read out by a pixie4 daq system @xcite with 14-bit adc s at a sampling rate of 75 mhz . the resolution ( fwhm ) of the core
is @xmath43.5 kev at 1.3 mev and those of the segments are between 2.5 and 4.0 kev .
a time resolution of roughly 10 ns can be achieved with the sampling rate used .
this corresponds to a position resolution of @xmath41 mm inside the detector volume.1 cm per 100 ns . ]
more information about the segmented detector and the daq system can be found in @xcite .
a 100 kbq @xmath8th source is positioned at a distance of 23@xmath91 cm from the center of the segmented detector and faces the center point of segment 14 , as illustrated in figure [ fig : setup ] . a second non - segmented and well - type germanium detector , a canberra reversed germanium detector ( rege ) @xcite , is positioned at the same height with the closed end facing the segmented germanium detector .
the distance from the closed end surface to the center of the segmented detector is 23@xmath91 cm .
the rege crystal is 60 mm in height and 65 mm in diameter .
it has a resolution ( fwhm ) of 2.3 kev at 1.3 mev .
it is used to tag the photons scattered mostly in segment 14 .
the geometrical acceptance of the rege detector results in recorded scs events with scattering angles between @xmath4 65@xmath6 and @xmath4 80@xmath6 corresponding to energy depositions in the segmented detector between @xmath4 1940 kev and @xmath4 2110 kev .
the precision of the alignment of the rege detector with respect to the @xmath8th source and the segmented detector is @xmath4 5@xmath6 .
the energy thresholds for all channels are set to 100 kev .
a coincidence trigger is required between the core of the segmented detector and the rege with a coincident time window of 500 ns . due to a technical limitation of the coincidence trigger of the daq system ,
only four channels could be read out .
thus , for each coincidence trigger , only the energies of the core ( @xmath10 ) , segment 14 ( @xmath11 ) , segment 17 ( @xmath12 ) ( below segment 14 , as illustrated in figure [ fig : setup ] ) and the rege ( @xmath13 ) were recorded .
300 time samples were taken for each pulse for all 4 channels .
this corresponds to a time window of 4 @xmath14s including 1 @xmath14s before the arrival of the trigger . in this analysis ,
however , only the core pulses are used for the psa .
the actual coincidence trigger rate was @xmath4 12 hz . the independent trigger rates of the segmented detector and of the rege detector were both @xmath42000 hz .
this results in an accidental coincidence rate of @xmath42 hz .
the coincidence trigger rate without the @xmath8th source is @xmath150.1 hz .
therefore , without further cuts , @xmath4 20% of all events are expected to originate from accidental coincidences .
however , the fraction of accidental coincidence events among the selected scs events is negligible , as discussed in the next section . in total 360 000 coincident events
were collected .
four different data samples are selected : * @xmath16 : single - compton - scattering ( scs ) events + @xmath17 5.0 kev + @xmath18 @xmath19 kev + @xmath18 @xmath20 3.0 kev * @xmath21 : events with the 2.6 mev photon fully absorbed in the segmented detector + @xmath22 5.0 kev * @xmath23 : dep events + @xmath24 5.0 kev ( two 511 kev photons escape . )
* @xmath25 : sep events + @xmath26 5.0 kev ( one 511 kev photon escapes . ) the @xmath16 sample is selected through three cuts .
the allowed window of @xmath95 kev of the sum energy of both detectors around 2614.5 kev covers about three times the combined energy resolution ( 3@xmath27 ) of the detectors .
the geometrical acceptance for scs events extends to 2110 kev , but sep events would contaminate the sample , as they have a core energy of @xmath10 = 2103.5 kev in this setup .
they are excluded by removing events with the core energy of the segmented detector above 2090 kev .
the @xmath5tl decay also produces 583.2 kev photons with a branching ratio of 84.5% . to avoid coincidences orginating from these photons an energy window of @xmath28
3.0 kev is excluded .
the single - segment events are selected from each data sample by additionally requiring * single - segment requirement : + @xmath29 kev @xmath30 @xmath31 kev the single - segment event samples are noted as @xmath32 , @xmath33 , @xmath34 and @xmath35 , respectively .
the coincidence trigger is only relevant for the @xmath16 sample .
however , the other samples are selected out of the collected coincident events to ensure the same experimental conditions . in principle
the rege detector could also be used to tag 511 kev photons for events in the @xmath25 and @xmath23 samples .
however , the statistics available is not sufficient .
the distribution of the energy of the core , @xmath10 , of all coincident events is shown in figure [ fig : energy_data_mc]a .
a ) for all coincident events b ) for events with @xmath10+@xmath13=(2614 @xmath9 5 ) kev .
the 8 selected samples are indicated .
the predicted distributions from the monte carlo are shown as well . ]
the @xmath10 distribution of all single - segment coincident events is shown in the same plot .
also shown are the simulated spectra which will be discussed in the next section .
figure [ fig : energy_data_mc]b shows the @xmath10 distribution for all coincident events with @xmath36 kev .
the arrows indicate the @xmath10 range corresponding to the acceptance angles for the @xmath16 sample .
.the numbers of events in all data samples are presented in the first row . for the @xmath16 sample ,
@xmath37 in the second row corresponds to the fraction of events with @xmath38 107 ns . for the @xmath21 , @xmath23 and @xmath25
, it corresponds to the fraction of events in the central peaks of the @xmath39 distributions .
the ratios of event numbers for data and mc are given in the third row with statistical errors only . [ cols="^,^,^,^,^,^,^,^,^",options="header " , ] the same procedure as described in the previous section is repeated with the ann trained using the @xmath16 ( @xmath40 ) as the sse - dominant and the @xmath21 ( @xmath41 ) as the mse - dominant samples .
the values of @xmath42 versus @xmath43 corresponding to the maximum slope are shown in figure [ fig : psa_dep_vs_sep]d .
the resulting identification probabilities @xmath44 and @xmath45 are given in the last two rows of table [ table : psa_results ] .
the ann can correctly identify sse like events at the 70 % and mse like events at the 80 % level .
this confirms again that the selected scs samples are enriched in sse like events and can be used to train the ann package .
events with photons compton scattering only once inside a germanium detector , scs events , can be selected by tagging the scattered photon with a second germanium detector .
the pulse shapes of these events can be studied and used to test methods that distinguish between single - site and multi - site events . in order to collect scs events and perform pulse shape analysis , an 18-fold segmented prototype detector for the phase - ii of the gerda experiment
was positioned in front of a @xmath8th source .
a second germanium detector was positioned to record the escaped photons at 72@xmath6 , corresponding to 2040 kev energy deposit in the segmented detector , close to the @xmath46-value of the 0@xmath2@xmath3 decay of @xmath1ge . according to the mc simulation
@xmath472 % of the collected scs events are true sse events .
the sse - dominance is verified by an artifical neural network ( ann ) trained in an independent way .
these scs events are then themselves used to train the pulse shape analysis package and thus the trained psa is able to identify single- and multi - site events with efficiencies at the @xmath480% level .
future studies can improve in two ways .
the fraction of sse events in the collected scs sample can be increased by further improving the tagging method .
for example , the whole experimental setup can be shielded from external photons and collimators can be positioned between the two detectors .
the monte carlo predictions can also be improved .
currently they are based on the size of the energy deposits only .
better predictions require a detailed pulse shape simulation which is currently being developed for the detectors under study . | compton scattering is one of the dominant interaction processes in germanium for photons with an energy of around two mev .
if a photon scatters only once inside a germanium detector , the resulting event contains only one electron which normally deposits its energy within a @xmath0 range .
such events are similar to @xmath1ge neutrinoless double beta - decay ( 0@xmath2@xmath3 ) events with just two electrons in the final state .
other photon interactions like pair production or multiple scattering can result in events composed of separated energy deposits .
one method to identify the multiple energy deposits is the use of timing information contained in the electrical response of a detector or a segment of a detector .
the procedures developed to separate single- and multiple - site events @xcite are tested with specially selected event samples provided by an 18-fold segmented prototype germanium detector for phase ii of the germanium detector array , gerda @xcite . the single compton scattering , i.e. single - site , events are tagged by coincidently detecting the scattered photon with a second detector positioned at a defined angle .
a neural network is trained to separate such events from events which come from multi - site dominated samples .
identification efficiencies of @xmath4 80% are achieved for both single- and multi - site events . , , , , , double beta decay , germanium detectors , pulse shape analysis 23.40.-s , 14.60pq , 29.40.-n |
many cosmological observations @xcite , such as the type ia supernova(sn ia ) , wilkinson microwave anisotropy probe(wmap ) , the sloan digital sky survey(sdss ) etc . , support that our universe is undergoing an accelerated expansion .
this accelerated expansion is always attributed to the dark energy with negative pressure which induces repulsive gravity 5,6 . except for the property of negative pressure
, we do not know too much about the components and other properties of the dark energy .
what is more , we have no idea about how the dark energy evolves .
various dark energy models have been proposed to describe the evolution , all of which must be constrained by astronomical observations . in these models ,
the equation of state @xmath2 plays a key role and can reveal the nature of dark energy which accelerates the universe , where @xmath3 and @xmath4 are the pressure and energy density of the dark energy respectively .
different equations of state lead to different dynamical changes and may influence the evolution of our universe . @xmath5 and
its time derivative with respect to hubble time @xmath6(where @xmath7 is scale factor ) are currently constrained by the distance measurements of sn ia , and the current observational data constrain the range of equation of state as @xmath8 @xcite . the cosmological constant model with equation of state @xmath9 , which is considered as the most acceptable candidate for dark energy has been investigated in @xcite .
the quintessence scalar field with @xmath5 in the range from @xmath10 to @xmath11 has been investigated in @xcite . the phantom scalar field with @xmath12 , which can introduce the negative kinetic energy and violate the well known energy condition , has been found in @xcite .
in order to get more information on dark energy , it has been widely discussed from the thermodynamical viewpoints , such as thermodynamics of dark energy with constant @xmath13 in the range @xmath14 @xcite , @xmath15 in the de sitter space - time and anti - de sitter space - time @xcite , @xmath16 in the phantom field @xcite and the generalized chaplygin gas @xcite and so on .
more discussions on the thermodynamics of dark energy can be found in @xcite . in this paper
, we investigate the thermodynamical properties of an accelerated expanding universe driven by dark energy with the equation of state @xmath0 and the hubble parameter @xmath17 $ ] ( where @xmath18 is the hubble constant and @xmath19 is the parameter of matter density ) .
this model is in agreement with the observations in low redshift , where the parameter @xmath20 and @xmath21 are suggested by @xcite .
though combining quantum mechanics with general relativity , it was discovered that a black hole can emit particles .
the temperature of a black hole is proportional to its surface gravity and the entropy is also proportional to its surface area @xcite .
the hawking temperature is given as @xmath22 , where @xmath23 is the surface gravity of a black hole , and the entropy of a black hole is @xmath24 , where @xmath25 is its surface area .
hawking temperature , black hole entropy , and the black hole mass satisfy the first law of thermodynamics @xmath26 @xcite . if a charged black hole is rotating , the thermodynamical first law is expressed as @xmath27 , where @xmath28 , @xmath29 , @xmath30 , @xmath31 are the dragged velocity , angular momentum , electric potential and charge of a black hole respectively .
the event horizon of a stationary black hole is considered as the system boundary @xcite , inside which the black hole should maintain thermodynamical equilibrium .
if our universe can be considered as a thermodynamical system@xcite , the thermodynamical properties of the black hole can be generalized to space - time enveloped by the apparent horizon or the event horizon .
the thermodynamical properties of the universe may be similar to those of the black hole , the thermodynamical laws should be satisfied .
this paper is organized as follows : in sec ii , we examine the thermodynamical first and the second laws on the apparent horizon with @xmath32 . in sec iii
, we similarly examine the thermodynamical laws on the event horizon . in sec iv
, we draw some conclusions and discussions .
the friedmann - robertson - walker metric is @xcite @xmath33 if setting @xmath34the dynamical apparent horizon can be determined by the relation@xmath35 so we can get the apparent horizon radius @xmath36 the wmap data suggest that our universe is spatially flat with @xmath37 , so the apparent horizon is @xmath38 in this case .
the surface gravity on the apparent horizon is @xmath39and the temperature is @xmath40 in this paper , we concentrate on a more general model with @xmath41 . \label{x2}\ ] ] as we just want to know more properties about dark energy , we take @xmath42 . and the hubble parameter becomes @xmath43 in this section we use temperature , entropy and energy to examine the first and the second laws of thermodynamics on the apparent horizon
. the temperature and entropy of the universe on the apparent horizon are @xmath44 and @xmath24 respectively . when calculating the flux of energy through the area @xmath45 of the apparent horizon , we treat the horizon as a static surface , but allowing it to vary slowly with time .
the amount of energy flux @xcite crossing the apparent horizon within the time interval @xmath46 is @xmath47the entropy of the apparent horizon is @xmath48 and its differential form is @xmath49 thus we can get @xmath50from eq.(9 ) and eq.(12 ) , we obtain the result @xmath51 , so the first law of thermodynamics on the apparent horizon of this redshift - dependent model is confirmed .
the entropy of the universe inside the apparent horizon is related to its energy and pressure through @xmath52the energy inside the apparent horizon is @xmath53 and the volume is @xmath54 .
so we get @xmath55 in which @xmath0 and @xmath56 .
thus we can get the derivative of @xmath57 with respect to @xmath1 @xmath58for the entropy of the apparent horizon , we get@xmath59and the derivation of @xmath60 with respect to @xmath61 is @xmath62 the total entropy varies with redshift @xmath61 , so its derivation is @xmath63 considering the relation @xmath64 , we get @xmath65 we find that the total entropy of the apparent horizon does not decrease with time and the second thermodynamical law is satisfied on the apparent horizon .
the event horizon of the universe is the position of the greatest distance that an particle can reach at a particular cosmic epoch , so the definition of the event horizon is @xmath66 the energy flux through the event horizon can be similarly expressed as @xmath67 meanwhile , we get @xmath68the entropy of the event horizon is@xmath69using hawking temperature of the event horizon , we can also obtain @xmath70 it is shown in fig.1 and fig.2 that @xmath71 in the higher redshift , @xmath72 is in agreement with @xmath73 , but in the lower redshift , they are not consistent .
therefore , the first law of thermodynamics with the usual definition of entropy and temperature on the event horizon is not always confirmed . for the entropy of matter and fluids inside the event horizon , we have@xmath74the energy inside the event horizon is @xmath75 , so we get @xmath76the entropy of the event horizon is @xmath77 , so we have @xmath78similarly we get @xmath79\frac{dr_{e}}{% dz}. \label{b7}\end{aligned}\ ] ] in fig.3
, we also find the total entropy inside the event horizon is not always negative . through the relation @xmath64 , we can get @xmath80as a result , @xmath81 does not keep positive all the time . the thermodynamical second law breaks down inside the event horizon .
the hawking temperature , entropy together with the black hole mass follow the thermodynamical first law @xmath26 , and the einstein equation can be derived from the thermodynamical law @xcite . as the thermodynamical laws hold on the event horizon of a stationary black hole , the black hole keeps thermodynamical equilibrium inside it .
the apparent horizon of a black hole is a two - dimensional surface for which the outgoing orthogonal null geodesics have zero divergence @xcite .
a black hole in asymptotically flat space - time is defined as a region so that no casual signal ( i.e. , a signal propagating at a velocity not greater than light ) from light can reach , then the event horizon constitutes the boundary of a black hole . when the weak energy condition is satisfied , the apparent horizon either coincides with event horizon or lies inside it . in stationary black holes ,
the apparent horizon and the event horizon coincide , such as schwarzschild black hole and kerr - newman black hole and so on . there always exists the event horizon located outside the apparent horizon .
when a spherically symmetric black hole becomes non - stationary for some time due to the matter falling into it , the thermodynamical laws are not fit for a non - stationary black hole . when a black hole is non - stationary , the apparent horizon is always not consistent with the event horizon and the thermodynamical laws may not always hold on the event horizon .
the universe is accelerating and should be not stationary . due to the heat through apparent horizon area , we have calculated a variation of the geometrical entropy on the apparent horizon .
we find that the thermodynamical relation is @xmath82 .
in addition , we also examine the changes of the total entropy with the time @xmath83 , therefore , the second law of thermodynamics is evidently confirmed .
while the thermodynamical laws on the event horizon do not hold on , the results are in agreement with the work of wang @xcite .
when the apparent horizon can be regarded as a thermodynamical system which has associated entropy @xmath84 with the temperature @xmath85 , the differential form of the first thermodynamical law on the apparent horizon can be written to the differential form of the friedmann equation @xcite . in a word ,
the apparent horizon is a good boundary of keeping thermodynamical properties , and the universe has a thermal equilibrium inside it .
these properties disappear on the event horizon .
maybe the usual definition of temperature and entropy are not well - defined , or it is non - equilibrium of thermodynamics inside the universe . in fig.4 , in the low redshift , the changing rate of the apparent horizon and the event horizon is not significant.the thermodynamical description of horizons @xcite will be approximately valid .
the apparent horizon always exists and usually is inside the event horizon . in fig.5 ,
the event horizon is outside apparent horizon at higher redshift phase and the universe is undergoing accelerated expansion .
but in this model the redshift becomes smaller than zero and the equation of state is @xmath86 the universe may have a superacceleratd expanding process and the apparent horizon is outside the event horizon .
this model is similar with the phantom field which will lead to undesirable future singularity and violate the weak energy condition . in summary
, it is interesting to investigate the universe from the thermodynamical viewpoints , although thermodynamical properties need to be explored more deeply in the future .
yongping zhang would like to thank shiwei zhou and xiaokai he for their valuable suggestions and help .
this work is supported by the national natural science foundation of china(grant no.10473002 ) , the national basic research program of china ( grant no .
2003 cb 716302 ) and the scientific research foundation for the returned overseas chinese scholars , state education ministry . | the thermodynamical properties of dark energy are usually investigated with the equation of state @xmath0 .
recent observations show that our universe is accelerating , and the apparent horizon and the event horizon vary with redshift @xmath1 . when definitions of the temperature and entropy of a black hole are used to the two horizons of the universe , we examine the thermodynamical properties of the universe which is enveloped by the apparent horizon and the event horizon respectively .
we show that the first and the second laws of thermodynamics inside the apparent horizon in any redshift are satisfied , while they are broken down inside the event horizon in some redshift .
therefore , the apparent horizon for the universe may be the boundary of thermodynamical equilibrium for the universe like the event horizon for a black hole . |
consider a class of quantum many - body problems such that each member of the class is characterized by @xmath4 particles of type 1 and @xmath5 particles of a different type 2 , and by the interactions , both pairwise and multiparticle , that operate among any subset of particles . with the interactions in place , all of the states and
all of the observables of any exemplar of the class are determined by the integers @xmath4 and @xmath5 .
specifically , for any observable @xmath2 defined for the class , there exists a physical mapping from @xmath6 to @xmath2 .
this notion is trivially extended to more than two type of constituents .
quite obviously , the nuclear many - body problem defines a class of this kind , with @xmath4 and @xmath5 taken as the numbers @xmath0 and @xmath1 of protons and neutrons in a nuclide .
other examples coming easily to mind : @xmath7he-@xmath8he clusters , binary and ternary alloys , etc .
approaches to calculation or prediction of the properties of individual systems belonging to such a class span a broad spectrum from pure _ ab initio _ microscopic treatments to phenomenological models having few or many adjustable parameters , with hybrid macroscopic / microscopic and density - functional methods in between .
these approaches are `` theory - thick '' to varying degrees , with the _ ab initio _ ones based in principle on exact theory and the phenomenological ones invoking physical intuition , semi - classical pictures , and free parameters .
thinking in the spirit of edwin jaynes , inventor of the maxent method and charismatic proponent of bayesian probability,@xcite it becomes of special interest to go all the way in the `` theory - thin '' direction and ask the question : * _ to what extent does the existing data on property @xmath2 across the members of a system class , _ and only the data _ , determine the mapping @xmath9 ? _ in general , this mapping takes one of two forms , depending on whether @xmath2 is a continuous variable ( e.g. , the nuclear mass excess or quadrupole moment ) or a discrete variable ( e.g. , the nuclear spin and parity ) .
the former case defines a problem of function approximation , while the latter defines a classification problem . during the past three decades
, powerful new methods have been developed for attacking such problems .
chief among these are advanced techniques of statistical learning theory , or `` machine learning , '' with artificial neural networks as a subclass . considering the concrete example of the mapping @xmath10 that determines the nuclear ( i.e. , atomic ) mass , a learning machine consists of ( i ) an input interface where @xmath0 and @xmath1 are fed to the device in coded form , ( ii ) a system of intermediate processing elements , and ( iii ) an output interface where an estimate of the mass appears for decoding . given a body of training data to be used as exemplars of the desired mapping ( consisting of input `` patterns , '' also called vectors , and their associated outputs ) , a suitable learning algorithm is used to adjust the parameters of the machine , e.g. , the weights of the connections between the processing elements in the case of a neural network .
these parameters are adjusted in such a way that the learning machine ( a ) generates responses at the output interface that reproduce , or closely fit , the masses of the training examples , and ( b ) serves as a reliable predictor of the masses of test nuclei absent from the training set .
this second requirement is a strong one the system should not merely serve as a lookup table for masses of known nuclei ; it should also perform well in the much more difficult task of prediction or _
generalization_. the most widely applied learning machine is the multilayer perceptron ( mlp ) , consisting of a feedforward neural network with at least one layer of `` hidden neurons '' between input and output interfaces.@xcite mlps are usually trained by the backpropagation algorithm,@xcite essentially a gradient - descent procedure for adjusting weight parameters incrementally to improve performance on a set of training examples . a significant measure of success has been achieved in constructing global models of nuclear properties based on such networks , with applications to atomic masses , neutron and proton separation energies , spins and parities of nuclear ground states , stability vs. instability , branching ratios for different decay modes , and beta - decay lifetimes .
( reviews and original references may be found in ref . 5 .
) the support vector machine ( svm),@xcite a versatile and powerful approach to problems in classification and nonlinear regression , entered the picture in the 1990s .
rooted in the strategy of structural - risk minimization,@xcite it has become a standard tool in statistical modeling .
although multilayer perceptrons as well as support vector machines are in principle universal approximators , svms eliminate much of the guesswork of mlps , since they incorporate an automatic process for determining the architecture of the learning machine . not surprisingly , they have become the method of choice for a wide variety of problems .
our selection of global nuclear systematics as a concrete example for the application of advanced machine - learning algorithms is neither accidental nor academic .
there exists a large and growing body of excellent data on nuclear properties for thousands of nuclides , providing the raw material for the construction of robust and accurate statistical models .
moreover , interest in this classic problem in nuclear physics has never been greater .
the advent of radioactive ion - beam facilities , and the promise of the coming generation epitomized by the rare isotope accelerator ( ria ) , have given new impetus to the quest for a unified , global understanding of the structure and behavior of nuclei across a greatly expanded nuclear chart .
the creation of hundreds of new nuclei far from stability opens exciting prospects for discovery of exotic phenomena , while presenting difficult challenges for nuclear theory . following the pattern indicated above , traditional methods for theoretical postdiction or prediction of the properties of known or unknown nuclides include _
ab initio _ many - body calculations employing the most realistic nuclear hamiltonians@xcite and , more commonly , density functional approaches and semi - phenomenological models . since computational barriers limit _ ab initio _ treatment to light nuclei ,
viable global models inevitably contain parameters that are adjusted to fit experimental data on certain reference nuclei .
global models currently representing the state of the art are hybrids of microscopic theory and phenomenology ; most notably , they include the macroscopic / microscopic droplet models of mller et al.@xcite and the density functional theories employing skyrme , gogny , or relativistic mean - field lagrangian parametrizations of self - consistent mean - field theory.@xcite from the standpoint of data analysis , these approaches are inherently theory - thick , since their formulation rests on a deep knowledge of the problem domain .
it is evident that data - driven , `` theory - thin , '' statistical models built with machine - learning algorithms can never compete with traditional global models in providing new physical insights .
nevertheless , in several respects they can be of considerable value in complementing the traditional methods , especially in the present climate of accelerated experimental and theoretical exploration of the nuclear landscape . *
a number of studies@xcite suggest that the quality and quantity of the data has already reached a point at which the statistical models can approach and possibly surpass the theory - thick models in sheer predictive performance . in this contribution , we shall present strong evidence from machine learning experiments with support vector machines that this is indeed the case . * in spite of their `` black - box '' origin , the machine - learning predictions can be directly useful to nuclear experimentalists working at radioactive ion - beam facilities , as well as astrophysicists developing refined models of nucleosynthesis . *
although not straightforward , it will in fact be possible to gain some insights into the inner workings of nuclear physics through statistical learning experiments , by applying techniques analogous to gene knock - out studies in molecular biology .
* it is fundamental interest to answer , for the field of nuclear physics , the jaynesian question that was posed above .
in technical jargon , the support vector machine is a _ kernel method_,@xcite which , in terms of a classification problem , means that it implicitly performs a nonlinear mapping of the input data into a higher - dimensional _ feature space _ in which the problem becomes separable ( at least approximately ) .
architecturally , the support vector machine and multilayer perceptron are close relatives ; in fact the svm includes the mlp with one hidden layer as a special case .
however , svms in general offer important advantages over mlps , including avoidance of the curse of dimensionality through extraction of the feature space from the training set , once the kernel and error function have been specified .
support vector machines may be developed for function approximation ( i.e. , nonlinear regression ) as well as classification . in either case , the output of the machine ( approximation to the function or location of the decision hyperplane , respectively ) is expressed in terms of a representative subset of the examples of the mapping @xmath11 contained in the training set .
these special examples are the _ support vectors_. the basic ideas underlying svms are most readily grasped by first considering the case of a classification problem involving linearly separable patterns .
suppose some of the patterns are green and the others are red , depending on some input variables defining the @xmath12-dimensional input space . to find a decision surface that separates red from green patterns ,
one _ seeks the hyperplane that provides the maximum margin between the red and green examples_. the training examples that define the margin are just the support vectors . in this simple case , an exact solution is possible , but in general errors are unavoidable . when faced with a problem involving nonseparable patterns , the objective then is to locate a decision hyperplane such that the misclassification error , averaged over the training set , is minimized .
guided by the principle of structural - risk minimization,@xcite the svm approach determines an optimal hyperplane by minimizing a cost function that includes a term to reduce the vc dimension@xcite ( thereby enhancing generalization capability ) and a term that governs the tradeoff between machine complexity and the number of nonseparable patterns . in practice ,
the svm strategy actually involves two steps .
the first is to implement a nonlinear mapping @xmath13 : @xmath14 , from the space of input vectors @xmath15 into a higher - dimensional feature space , which is `` hidden '' from input and output ( and corresponds to the hidden layer in mlps ) .
this is done in terms of an inner - product kernel @xmath16@xmath13@xmath17@xmath13@xmath18
satisfying certain mathematical conditions , notably mercer s theorem.@xcite the second step is to find a hyperplane that separates ( approximately , in general ) the features identified in the first step .
this is accomplished by the optimization procedure sketched above .
a self - contained introduction to the svm technique is beyond the scope of the present contribution .
excellent treatments are available in the original work of vapnik as expounded in refs .
6,7 and in haykin s text@xcite ( see also ref .
19 ) . to provide some essential background ,
let us consider a regression problem corresponding to a map @xmath19 , where @xmath15 is an input vector with @xmath12 components @xmath20 , and suppose that @xmath21 training examples indexed by @xmath22 are made available .
then the optimal approximating function takes the form @xmath23 solution of the optimization problem stated above determines the parameters @xmath24 and @xmath25 , and the support vectors of the machine are defined by those training patterns for which @xmath26 .
different choices of the inner - product kernel appearing in eq .
( [ est ] ) yield different versions of the support vector machine .
common choices include @xmath27 corresponding to the _ polynomial learning machine _ with user - selected power @xmath28 ; a gaussian form @xmath29 containing a user - selected width parameter @xmath30 , which generates a radial - basis - function ( rbf ) network ; and @xmath31 which realizes a two - layer ( one - hidden - layer ) perceptron , only one of the parameters @xmath32 , @xmath33 being independently adjustable .
we also draw attention to a generalization of the rbf kernel ( [ rbf ] ) introduced recently as a simplified version of what is called anova decomposition,@xcite having the form @xmath34 \right)^d \ , , \label{anova}\ ] ] the support vector machine may be considered as a feedforward neural network in which the inner - product kernel , through an appropriate set of @xmath35 elements @xmath36 , defines a layer of hidden units that embody the mapping from the @xmath12-dimensional input space to the @xmath35-dimensional feature space .
these hidden units process the input patterns nonlinearly and provide outputs that are weighted linearly and summed by an output unit . as already pointed out , the familiar structures of radial - basis - function networks and two - layer perceptrons can be recaptured as special cases by particular choices of kernel .
however , the svm methodology transcends these limiting cases in a very important way : it automatically determines the number of hidden units suitable for the problem at hand , whatever the choice of kernel , by finding an optimally representative set of support vectors and therewith the dimension of the feature space .
in essence , the support vector machine offers a generic and principled way to control model complexity .
by contrast , approaches to supervised learning based on mlps trained by backpropagation or conjugate - gradient algorithms depend heavily on rules of thumb , heuristics , and trial and error in arriving at a network architecture that achieves a good compromise between complexity ( ability to fit ) and flexibility ( ability to generalize ) .
in this section we summarize the findings of recent explorations of the potential of support vector machines for global statistical modeling of nuclear properties .
the discussion will focus on the predictive reliability of svm models relative to that of traditional `` theory - thick '' models .
the properties that are directly modeled in these initial studies , all referring to nuclear ground states , are ( i ) the nuclear mass excess @xmath37 , where @xmath38 is the atomic mass , measured in amu , ( ii ) @xmath39-decay lifetimes of nuclides that decay 100% via the @xmath40 mode , and ( iii ) nuclear spins and parities . the requisite experimental data are taken from the on - line repository of the brookhaven national nuclear data center ( nndc ) at http://www.nndc.bnl.gov/. the experimental mass values are those of the ame03 compilation of audi et al.@xcite extensive preliminary studies have been performed to identify inner - product kernels well suited to global nuclear modeling .
earlier work converged on the anova kernel ( 5 ) as a favorable choice , and corresponding results have been published in ref . 19 .
more recently , we have introduced a new kernel that yields superior results , formed by the sum of polynomial and anova kernels and named the pa kernel .
( satisfaction of mercer s theorem is conserved under summation . )
the new kernel contains three parameters ( @xmath28 , @xmath30 , and @xmath41 ) that may be adjusted by the user . aside from parameters contained in the inner - product kernel ,
the svm procedure involves a constant @xmath42 giving the user control over the tradeoff between complexity and flexibility , plus an additional control constant @xmath43 in the regression case , measuring the tolerance permitted in the reproduction of training data .
thus , svm models developed with the pa kernel contain four or five adjustable parameters ( five in all applications reported here ) . to allow for a meaningful evaluation of predictive performance ( whether interpolation or extrapolation ) , the existing database for the property being modeled is divided into three subsets , namely the _ training set _
, _ validation set _ , and _
test set_. these sets are created by random sampling , consistently with approximate realization of chosen numerical proportions among them , e.g. ( 1002@xmath44):@xmath44:@xmath44 for training , validation , and test sets , respectively , with @xmath45 . the training set is used to find the support vectors and construct the machine for given values of the adjustable parameters @xmath28 , @xmath30 , @xmath41 , @xmath42 , and @xmath43 .
the validation set is used to guide the optimal determination of these parameters , seeking good performance on both the training and validation examples .
the test set remains untouched during this process of model development ; accordingly , the overall error ( or error rate ) of the final model on the members of the test set may be taken as a valid measure of predictive performance .
when one considers how svm models might be applied in nuclear data analysis during the ongoing exploration of the nuclear landscape , it seems reasonable that consistent predictive performance for 80:10:10 or 90:5:5 partitions into training , validation , and test sets would be sufficient for the svm approach to be useful in practice .
the svm approach has been applied to generate a variety of global models of nuclear mass excess , beta - decay lifetimes , and spin / parity , corresponding to different kernels , databases , partitions into training / validation / test sets , and tradeoffs between the relative performance on these three sets . here
we will focus on those models considered to be the best achieved to date .
moreover , due to limited space , we will restrict the discussion to the most salient features of those models and to an assessment of their quality relative to favored traditional global models and to the best available mlp models .
further , more detailed information may be found at the web site http://abacus.wustl.edu/clark/svmpp.php , which generates svm estimates of the listed nuclear properties for @xmath46 pairs entered by visitors .
this web site will be periodically updated as improved svm models are developed .
development and testing of the svm mass models to be highlighted here are based on the ame03 data for all nuclides with @xmath47 and experimental masses having error bars below 4% .
this set of nuclides is divided into the four classes : even-@xmath0-even-@xmath1 ( ee ) even-@xmath0-odd-@xmath1 ( eo ) , odd-@xmath0-even-@xmath1 ( oe ) , and odd-@xmath0-odd-@xmath1 ( oo ) .
separate svm regression models were constructed for each such `` even - oddness '' class .
this does introduce some minimal knowledge about the problem domain into the modeling process ; one might therefore say that the models developed are not absolutely theory - free .
however , the data itself gives strong evidence for the existence of different mass surfaces depending on whether @xmath0 and @xmath1 are even or odd .
knowledge of the integral character of @xmath0 and @xmath1 may , quite properly , bias the svm toward inclusion of associated quantum effects.@xcite table 1 displays performance measures for models based on an 80:10:10 target partitioning of the full data set among training , validation , and test sets , respectively .
inspection of the actual distributions of these sets in the @xmath48 plane shows that substantial fractions of the validation and test sets lie on the outer fringes of the known nuclei , significantly distant from the line of stable nuclides .
accordingly , performance on the test set measures the capability of the models in extrapolation as well as interpolation .
performance on a given data set is quantified by the corresponding root - mean - square ( rms ) error @xmath49 of model results relative to experiment ( as is standard in global mass modeling ) . the `` optimized '' model parameters
are included in table 1 .
they show enough differences from one even - oddness class to another to justify development of separate models for the four classes .
the results in table 1 attest to a quality of performance , in both fitting and prediction , that is on a par with the best available from traditional modeling@xcite and from mlp models trained by an enhanced backpropagation algorithm.@xcite to emphasize this point qualitatively , we display in table 2 some representative rms error figures that have been achieved in recent work with all three approaches .
( we must note , however , that the data sets used for the different entries in the table may not be directly comparable , and the division into training , validation , and test sets does not necessarily have the strict meaning assigned here . )
the second svm model listed in the table was developed for a partitioning of the data into training , validation , and test sets of approximately 90:5:5 , obtained by random transfer of nuclides from the validation and test sets of the 80:10:10 model to the training set .
the quality of representation that can be realized through the svm methodology may be highlighted in another way .
employing the nuclear mass excess values generated by the svm models of table 1 , we have calculated the @xmath50 values for eight alpha - decay chains of the superheavy elements 110 , 111 , 112 , 114 , 115 , 116 , and 118 .
( the alpha - decay @xmath51-value is defined as @xmath52 , where be stands for the binding energy of the indicated nuclide . )
results are presented in graphical and tabular form on the web site http://haochen.wustl.edu/svm/svmpp.php .
for the models of table 1 ( based on an 80:10:10 partition of the assumed ame03 data set ) , the average rms error of the 38 estimates of @xmath50 is 0.82 mev , while the average absolute error is 0.64 mev . we emphasize that these estimates are predictions ( rather than fits ) , since none of the nuclei involved belongs to the validation or test set .
moreover , due to the situation of these superheavy nuclides in the @xmath48 plane , prediction of the associated @xmath50 values provides a strong test of extrapolation .
the performance of svm mass models documented in tables 1 and 2 and in the alpha - chain predictions gives assurance that this approach to global modeling will be useful in guiding further exploration of the nuclear landscape .
however , it is important to gain some sense of when and how it begins to fail .
the performance figures for the two sets of svm models involved in table 2 are consistent with the natural expectation that if one depletes the validation and test sets of the 80:10:10 partition in favor of an enlarged test set , the predictive ability of the model is enhanced .
conversely , one should be able to `` break '' the svm modeling approach by random depletion of the training set of the 80:10:10 model in favor of larger validation and test sets .
eventually the training set will become too small for the method to work at all .
the results of a quantitative study of this process are shown in figure 1 .
writing the generic partition as ( 1002@xmath44):@xmath44:@xmath44 , the error measure increases roughly linearly with @xmath44 for @xmath44 greater than 10 .
in addition to direct statistical modeling using either svms or mlps , a promising hybrid approach is being explored .
recently , the _ differences _ @xmath53 between experimentally measured masses and the corresponding theoretical masses given by the finite - range droplet model ( frdm ) of mller , nix , and collaborators@xcite have been modeled with a feedforward neural network of architecture 46661 trained by a modified backpropagation learning algorithm.@xcite ( the integers denote the numbers of neurons in successive layers , from input to output . )
the rms errors on training ( 1276 ) , validation ( 344 ) , and test ( 529 ) sets are respectively 0.40 , 0.49 , and 0.41 mev , where the numbers of nuclides in each of these sets is given in parentheses . in a similar experiment , we have constructed svm models for ee , eo , oe , and ee classes using pa kernels
. overall rms errors of 0.19 , 0.26 , and 0.34 were achieved on the training ( 1712 ) , validation ( 213 ) , and test ( 213 ) sets , respectively , with little variation over even - oddness classes .
error figures over comparable subsets for the frdm model in question run around 0.7 mev , again with relatively little variation from subset to subset .
these results suggest that mlps and svms are capable of capturing some 1/2 to 2/3 of the physical regularities missed by the frdm . it remains to be seen whether the residual error has a systematic component or instead reflects a large number of small effects that will continue to elude global description .
another important problem in global modeling involves the prediction of beta - decay halflives of nuclei . as in the case of atomic masses , this is a problem in nonlinear regression .
here we restrict attention to nuclear ground states and to nuclides that decay 100% through the @xmath40 mode .
for this presentation , we make the further restriction to nuclides with halflives @xmath54 below @xmath55 s ( although we have also included the longer - lived examples in another set of modeling experiments ) .
the brookhaven nndc provides 838 examples fitting these criteria .
since the examples still span 9 orders of magnitude in @xmath54 , it is natural to work with @xmath56 and seek an approximation to the mapping @xmath57 in the form of svms .
again , we construct separate svms for the ee , eo , oe , and oo classes , and again a kernel of type pa is adopted .
the full data set is divided by random distribution into training , validation , and test sets in approximately the proportions 80:10:10 .
the performance of the favored models is quantified in table 3 .
here we use two measures to assess the accuracy of the svm results for training , validation , and test nuclides .
these are the rms error , again denoted by @xmath49 , and the mean absolute error @xmath58 of the model estimates of @xmath59 , relative to experiment .
detailed studies of beta - decay systematics within the established framework of nuclear theory and phenomenology include those of staudt et al.@xcite ( 1990 ) , hirsch et al.@xcite ( 1993 ) , homma et al.@xcite ( 1996 ) , and mller et al.@xcite ( 1997 ) .
however , comparison of the performance of the svm models with that of the models resulting from these studies is obscured by the differences in the data sets involved .
most significantly , the data set employed here is considerably larger than those used previously , including as it does many new nuclides far from stability .
analysis of svm performance on subsets of the data set , now in progress , will yield useful information on the efficacy of svm models relative to the more traditional ones , as we continue to develop improved global models of beta - decay systematics . on the other hand , mlp models for the beta - halflife problem
have been generated for the same data set as used in our svm study , allowing a meaningful comparison to be made .
the best mlp models created to date@xcite show values for the rms error @xmath49 over all even - oddness classes of 0.55 in training , 0.61 in validation , and 0.64 in prediction .
these values are somewhat larger than those seen in table 3 .
however , it must be pointed out that the mlp results were obtained with a smaller training set , so the efficacy of the two statistical methods appears to be about equal at this stage of development . that being the case , it is relevant to note that the recent mlp models represent a distinct advance earlier versions,@xcite and that those earlier statistical models already showed better performance over short - lived data sets than the conventional models of homma et al.@xcite and mller et al.@xcite the applications to prediction of atomic masses and beta - decay lifetimes demonstrate the predictive power of svms in two important problems of global nuclear modeling that involve function estimation .
the final two applications will probe the performance of svms in global modeling of the discrete nuclear properties of parity and spin .
in essence , these are problems of classification : `` which of a finite number of exclusive possibilities is associated with or implied by a given input pattern ? ''
support vector machines were first developed to solve classification problems , and good svm classifier software is available on the web.@xcite however , for convenience and uniformity we prefer to treat the parity and spin problems with the same svm regression technique as in the other examples , also using the pa choice of inner - product kernel . in the parity problem
, the decision of the regression svm is interpreted to be positive parity [ negative parity ] if the machine s output is positive [ negative ] . in the spin problem ,
the spin assigned by the machine is taken to be correct if and only if the numerical output ( after rescaling ) is within @xmath60 of the correct value ( in @xmath61 units ) .
as before , all data are taken from the brookhaven site .
for parity and spin , it is especially natural to create separate svm models for the different even - oddness classes .
however , as is well known , all ee nuclei have spin / parity @xmath62 . modeling
this property is trivial for svms , so the ee class may be removed from further consideration .
the data in and of itself permits us to do so .
moreover , in the case of spin , the data itself establishes , with a high degree of certainty , that the spin of eo or oe nuclides takes half - odd integral values ( in units of @xmath61 ) , while the spin of oo nuclides is integral .
although this formulation of the parity and spin problems introduces significant domain knowledge into the model - building process , the data alone provides adequate motivation .
nuclei with spin values larger than 23/2 were not considered .
the predictive performance that may be achieved with svm models of parity and spin is illustrated in tables 4 and 5 .
performance is measured by the percentage of correct assignments .
construction of both parity and spin models is based on an 80:10:10 partition of the data into training , validation , and test sets .
( as usual , the target distribution is realized only approximately . )
averaged over even - oddness classes , the overall performance of the parity svms is 97% correct on the training set and 95% on the validation set , with a predictive performance on the test set of 94% .
obviously , assigning parity to nuclear ground states is an extremely easy task for support vector machines .
one might expect quite a different situation for the spin problem : since there are 12 legitimate spin assignments for the eo or oe nuclides considered ( i.e. , obeying the rules for addition of angular momenta ) and also 12 for the oo class , the chance probability of a correct guess is low .
it is then most remarkable that the svm spin models we have developed perform with very high accuracy in prediction as well as fitting and validation .
while some success has been achieved previously in mlp modeling of parity and spin,@xcite consistent predictive quality within the 8090 percentile range has been elusive . within main - stream nuclear theory and phenomenology ,
the problem of global modeling of ground - state spins has received little attention , and the few attempts have not been very successful . as a baseline , global nuclear structure calculations within the macroscopic / microscopic approach@xcite
reproduce the ground - state spins of odd-@xmath63 nuclei with an accuracy of 60% ( agreement being found in 428 examples out of 713 ) .
it should be mentioned that in the preliminary investigations described in ref .
19 , the tasks of global modeling of parity and spin with svms were in fact treated as classification rather than function - estimation problems .
corresponding svm classifiers were created using established procedures.@xcite based on an rbf kernel , results were obtained that surpass the available mlp models in quality , but are inferior to those reported here in tables 4 and 5 .
global statistical models of atomic masses , beta - decay lifetimes , and nuclear spins and parities have been constructed using the methodology of support vector machines .
the predictive power of these `` theory - thin '' models , which in essence are derived from the data and only the data , is shown to be competitive with , or superior to , that of conventional `` theory - thick '' models based on nuclear theory and phenomenology .
conservative many - body theorists may be troubled by the `` black - box '' nature of the svm predictors , i.e. , the impenetrability of their computational machinery . however ,
this alternative , highly pragmatic approach may represent a wave of the future in many fields of science
already visible in the proliferation of density - functional computational packages for materials physics and eventually molecular biology , which , for the user , are effectively black boxes . while it is true that the statistical models produced by advances in machine learning do not as yet yield the physical insights of traditional modeling approaches , their prospects for revealing new regularities of nature are by no means sterile .
this research has received support from the u.s . national science foundation under grant no .
we acknowledge helpful discussions and communications with s. athanassopoulos , m. binder , e. mavrommatis , t. papenbrock , s. c. pieper , and r. b. wiringa . in the regression studies
reported herein , we have found the mysvm software and instruction manual created by s. rping@xcite ( dortmund ) to be very useful . e. t. jaynes , _ probability theory : the logic of science _ ( cambridge university press , cambridge , 2003 ) .
s. haykin , _ neural networks : a comprehensive foundation _ , second edition ( mcmillan , new york , 1999 ) .
d. e. rumelhart , g. e. hinton , and r. j. williams , in _ parallel distributed processing : explorations in the microstructure of cognition _ ,
vol . 1 , edited by d. e. rumelhart _ et al . _
( mit press , cambridge , ma , 1986 ) . j. hertz , a. krogh , and r. g. palmer , _ introduction to the theory of neural computation _ ( addison - wesley , redwood city , ca , 1991 ) .
j. w. clark , t. lindenau , and m. l. ristig , _ scientific applications of neural nets _
( springer - verlag , berlin , 1999 ) .
v. n. vapnik , _ the nature of statistical learning theory _ ( springer - verlag , new york , 1995 ) .
v. n. vapnik , _ statistical learning theory _ ( wiley , new york , 1998 ) .
s. c. pieper and r. b. wiringa , _ annu .
_ * 51 * , 53 ( 2001 ) .
p. mller and j. r. nix , _
j. phys .
g _ * 20 * , 1681 ( 1994 ) .
p. mller , j. r. nix , w. d. myers , and w. j. swiatecki , _ at .
data nucl .
data tables _ * 59 * , 185 ( 1995 ) .
m. samyn , s. goriely , p .- h .
heenen , j. m. pearson , and f. tondeur , _ nucl .
phys . _ * a700 * , 142 ( 2002 ) ; s. goriely , m. samyn , p .- h .
heenen , j. m. pearson , and f. tondeur , _ phys .
c _ * 66 * , 024326 ( 2002 ) .
m. bender , p .- h .
heenen , and p .-
reinhard , _ rev .
* 75 * , 121 ( 2003 ) ; m. bender , g. f. bertsch , and p .- h .
heenen , _ phys .
_ * 94 * , 102503 ( 2005 ) .
g. f. bertsch , b. sabbey , and m. uusnkki , _ phys .
_ c * 71 * , 054311 ( 2005 ) .
k. a. gernoth and j. w. clark , _ neural networks _ * 8 * , 291 ( 1995 ) .
j. w. clark , e. mavrommatis , s. athanassopoulos , a. dakos , and k. a. gernoth , _ fission dynamics of atomic clusters and nuclei _ , edited by d. m. brink , f. f. karpechine , f. b. malik , and j. da providencia ( world scientific , singapore , 2001 ) , p. 76 .
[ nucl - th/0109081 ] s. athanassopoulos , e. mavrommatis , k. a. gernoth , and j. w. clark , _ nucl .
_ * a743 * , 222 ( 2004 ) , and references therein .
s. athanassopoulos , e. mavrommatis , k. a. gernoth , and j. w. clark , in _ advances in nuclear physics , proceedings of the hellenic symposium on nuclear physics _
, in press ( 2005 ) .
[ nucl - th/0509075 ] s. athanassopoulos , e. mavrommatis , k. a. gernoth , and j. w. clark , in _ advances in nuclear physics _ , proceedings of the hellenic symposium on nuclear physics , in press ( 2006 ) .
[ nucl - th/0511088 ] h. li , j. w. clark , e. mavrommatis , s. athanassopoulos , and k. a. gernoth , in _ condensed matter theories _ , vol .
20 , edited by j. w. clark , r. m. panoff , and h. li ( nova science publishers , hauppauge , ny , 2006 ) , p. 505 .
[ nucl - th/0506080 ] j. mercer , _ transactions of the london philosophical society ( a ) _ * 209 * , 415 ( 1909 )
. m. o. stitson , a. gammerman , v. vapnik , v. vovk , c. watkins , and j. weston , in _ advances in kernel methods support vector learning _ , edited by b. schkopf , c. burges , and a. j. smola ( mit press , cambridge , ma , 1999 ) , p. 285 .
g. audi , a. h. wapstra , c. thibault , j. blachot , and o. bersillon _ nucl .
phys . _ * a729 * ( 2003 ) .
s. gazula , j. w. clark , and h. bohr , _ nucl .
phys . _ * a540 * , 1 ( 1992 ) . s. athanassopoulos , e. mavrommatis , k. a. gernoth , and j. w. clark , to be published .
a. staudt , e. bender , k. muto , and h. v. klapdor - kleingrothaus , _ at .
data nucl .
data tables _ * 44 * , 132 ( 1990 ) .
m. hirsch , a. staudt , k. muto , and h. v. klapdor - kleingrothaus , _ at .
data nucl .
data tables _ * 53 * , 165 ( 1993 ) .
h. homma , e. bender , m. hirsch , k. muto , and h. v. klapdor - kleingrothaus , _ phys .
c _ * 54 * , 2972 ( 1996 ) .
p. mller , j. r. nix , and k. l. kratz , _ at .
data nucl .
data tables _ * 66 * , 131 ( 1997 ) .
n. costiris , a. dakos , e. mavrommatis , k. a. gernoth , and j. w. clark , to be published .
t. joachims ( 2004 ) , multi - class support vector machine , http://www.cs.cornell.edu/people/tj/svm_light/svm_multiclass.html ( 2004 ) .
j. w. clark , s. gazula , k. a. gernoth , j. hasenbein , j. s. prater , and h. bohr , in _ recent progress in many - body theories _ , vol . 3 , edited by t. l. ainsworth , c. e. campbell , b. e. clements , and e. krotscheck ( plenum , new york , 1992 ) , p. 371 .
k. a. gernoth , j. w. clark , j. s. prater , and h. bohr , _ phys .
_ * b300 * , 1 ( 1993 ) .
p. mller and j. r. nix , _ nucl .
phys . a _
* 520 * , 369c ( 1990 ) .
s. rping , mysvm , http://www-ai.cs.uni-dortmund.de/software/mysvm/ ( 2004 ) . | advances in statistical learning theory present the opportunity to develop statistical models of quantum many - body systems exhibiting remarkable predictive power .
the potential of such `` theory - thin '' approaches is illustrated with the application of support vector machines ( svms ) to global prediction of nuclear properties as functions of proton and neutron numbers @xmath0 and @xmath1 across the nuclidic chart . based on the principle of structural - risk minimization ,
svms learn from examples in the existing database of a given property @xmath2 , automatically and optimally identify a set of `` support vectors '' corresponding to representative nuclei in the training set , and approximate the mapping @xmath3 in terms of these nuclei .
results are reported for nuclear masses , beta - decay lifetimes , and spins / parities of nuclear ground states .
these results indicate that svm models can match or even surpass the predictive performance of the best conventional `` theory - thick '' global models based on nuclear phenomenology . |
let @xmath0 be a real bivariate polynomial and @xmath1 a one - form on @xmath3 .
let @xmath4 denote a continuous family of real ovals .
consider the _
abelian integral _
@xmath5 the _ infinitesimal _ hilbert sixteenth problem calls for the study of the zero set @xmath6 in particular , the goal is to obtain an upper bound @xmath7 on @xmath8 depending solely on the degrees of @xmath0 and @xmath1 . here and in the rest of the paper @xmath9 denotes the number of isolated points in the set @xmath10 .
the infinitesimal hilbert problem is motivated by the study of _ limit cycles _ born from the perturbation @xmath11 of the hamiltonian system @xmath12 .
in particular , the existence of the uniform bound @xmath7 may be seen as a particular case of the general hilbert sixteenth problem .
we refer the reader to the surveys @xcite for further details and references . the first _
general _ result concerning the infinitesimal hilbert problem was given in @xcite : [ thm : vk ] @xmath13 in other words , the number of zeros of abelian integral is uniformly bounded in terms of the degrees of the hamiltonian @xmath0 and the form @xmath1 .
however , this result is purely existential and does not give an explicit bound for @xmath14 .
the following uniform upper bound , established in @xcite , constitutes an explicit solution for the infinitesimal hilbert problem .
[ thm : bny ] @xmath15 where @xmath16 denotes an explicit polynomial of degree not exceeding 61 .
we call the attention of the reader to the fact that the dependence of the upper bound is _ doubly - exponential _ in both @xmath17 and @xmath2 .
in contrast , petrov and khovanskii proved the following result in an unpublished work ( see @xcite for an exposition ) .
[ thm : pk ] @xmath18 where @xmath19 , with @xmath20 some explicit function and @xmath21 _ some _ function of @xmath0 ( for which a bound is not given ) .
the bound given by theorem [ thm : pk ] is not uniform over the class of hamiltonians of a given degree , due to the appearance of the term @xmath22 .
however , using the methods developed in the proof of theorem [ thm : vk ] it is possible to prove that this term is in fact uniformly bounded @xcite .
[ thm : pk - z ] @xmath23 where @xmath19 , with @xmath20 some explicit funcition and @xmath24 _ some _ function of @xmath25 ( for which a bound is not given ) .
in summary , theorem [ thm : bny ] establishes an _ explicit _ bound on @xmath14 depending _ doubly - exponentially _ on @xmath26 , whereas theorem [ thm : pk - z ] establishes an _ existential _ bound depending _ linearly _ on @xmath26 .
the goal of this paper is to apply a combination of the ideas used in the proofs of these two results , to obtain an _ explicit _ bound depending _ linearly _ on @xmath26 .
the result is as follows .
we introduce the following notation to simplify the presentation of the results .
we write @xmath27 as a shorhand for @xmath28 .
we write @xmath29 for @xmath30 , and @xmath31 for @xmath32 .
finally we allow compositional iteration in the usual way , so @xmath33 corresponds to @xmath34 , etc .
[ thm : main - ai ] @xmath35 see subsection [ sec : concluding - remarks ] for a discussion of the percise form of the bound and possible improvements .
in this section we review background from the theory of analytic differential equations , the theory of abelian integrals , and the work @xcite .
let @xmath36 denote an @xmath37 rational matrix one - form over @xmath38 with singular locus @xmath39 .
the form is said to be _ integrable _ if @xmath40 .
this condition is equivalent to the existence of a fundamental solution matrix @xmath41 , defined over @xmath42 and ramified over @xmath39 , for the following system of equations @xmath43 in other words , we view @xmath44 as the matrix form of a connection defined on the trivial @xmath45-dimensional vector bundle over @xmath38 and @xmath46 as a fundamental matrix of horizontal sections .
let @xmath47 denote an affine chart on @xmath38 , and for convenience of notation let @xmath48 .
then the system may be viewed as a family of linear systems of differential equations in the @xmath49 variable parameterized by @xmath50 , @xmath51 we remark that not every system of the form may be obtained in this manner . in particular , systems obtained in this manner are necessarily _
isomonodromic_. the system is said to be _ regular _ if for any germ of a _ real analytic _
path @xmath52 with @xmath53 , the rate of growth of the fundamental solution matrix along @xmath54 is polynomial .
explicitly , we require that for suitable positive constants @xmath55 we have @xmath56 the analyticity of the curve @xmath54 is required to rule out spiralling around the singular locus . to each closed loop
@xmath57 one may associate a continuation operator @xmath58 describing the result of analytic continuation of @xmath41 along @xmath54 . the _ monodromy matrix _
@xmath59 is defined by the equation @xmath60 .
it is clear that @xmath61 depends only on the pointed homotopy class of @xmath54 , and that the conjugacy class of @xmath61 depends only on the free homotopy class of @xmath54 . in the future
we shall mainly be interested in the conjugacy class of the monodromy , and refer to the monodromy associated with a homotopy class of a closed loop in this sense .
a matrix @xmath62 is said to be _ quasi - unipotent _ if all of its eigenvalues are roots of unity .
equivalently , @xmath62 is quasi - unipotent if and only if there exist @xmath63 such that @xmath64 , where @xmath65 denotes the identity matrix .
we shall say that the monodromy along a loop @xmath54 is quasi - unipotent if the associated monodromy matrix @xmath61 is quasi - unipotent ( note that this condition depends only on the conjugacy class of @xmath61 ) . a loop @xmath54 is said to be a _ small loop _ around @xmath66 if there exists a germ of an analytic curve @xmath67 with @xmath68 such that @xmath54 is homotopic to a closed path @xmath69 for sufficiently small @xmath70 .
we shall only be interested in the case @xmath71 .
the system is said to be _ quasiunipotent _ if the monodromy matrix associated to each _ small loop _ is quasi - unipotent .
note that this condition does _ not _ imply that _ every _ monodromy matrix associated with the system is quasi - unipotent .
in particular , monodromies along loops encircling several singualities are often not small , and are not required to be quasi - unipotent ( and this is indeed the case in natural examples ) . in this subsection
we give definitions for measuring the complexity of the formulas representing various algebraic objects .
it is rather unusual in mathematics to be concerned with the particular formulas used for the description of an object .
questions of this form fall more neatly within the framework of _ mathematical logic_. indeed , strictly speaking the definitions in this subsection could be more accurately expressed in terms of logical formula complexity . in the interest of simplicity
we content ourselves with simple algebraic approximations of these notions which are sufficient for our purposes .
we stress that all definitions in this subsection refer to a particular representation of a given object . for instance , @xmath72 and @xmath73
are viewed as distinct fractional representations of the same polynomial .
a polynomial @xmath74 $ ] is said to be a _ lattice _ polynomial .
we shall say that such a polynomial is _ defined over @xmath75 _ , if @xmath76 where @xmath77 denotes a multiindex . we define the size of @xmath78 to be @xmath79 . a rational function given by a fraction of the form @xmath80
is said to be defined over @xmath75 if @xmath78 and @xmath81 are defined over @xmath75 . in this case , we define the size @xmath82 to be @xmath83 .
similarly , a one - form @xmath1 is said to be defined over @xmath75 if it is of the form @xmath84 where @xmath85 are rational functions defined over @xmath75 . in this case
, we define the size @xmath86 to be @xmath87 .
finally , say that a vector or a matrix is defined over @xmath75 if its of its components are , and define the size to be the sum of the sizes of components .
recall that we may view the system as a family of differential equations in the variable @xmath49 , of the form .
we shall be interested in studying the oscillatory behavior of the solutions of this equation .
however , due to the fact that the solutions of may be ramified , some care is required in measuring this oscillation .
let @xmath88 be a ( possibly multivalued ) function defined in a domain @xmath89 .
if @xmath90 is simply connected , then we define the following _ counting function _ as a measure for the number of zeros of @xmath88 : @xmath91 where @xmath92 varies over the branches of @xmath88 in @xmath90 ( which are well defined univalued functions , since @xmath90 is simply connected ) . for general domains , we use the following counting function , @xmath93 where @xmath94 varies over all triangular domains ( i.e. , domains whose boundary consists of straight line segment ) .
the restriction on the geometry of @xmath94 is needed in order to avoid spiralling around a singular point .
we stress that the closure of @xmath94 _ need not _ be contained in @xmath90 .
the boundary may contain singular points .
when @xmath90 is omitted from the notation , it is understood to be the domain of analyticity of the function @xmath88 .
let @xmath95 be a linear space of ( possibly multivalued ) functions defined in a domain @xmath89 . as a measure for the number of zeros of an element of @xmath95 , we use the following , @xmath96 when @xmath90 is omitted from the notation , it is understood to be the common domain of analyticity of the elements of @xmath95 .
[ rem : semicontinuity ] as remarked in @xcite , the counting function @xmath97 is lower semicontinuous with respect to the space @xmath95 . in particular , if we have a family of spaces @xmath98 continuously depending on a paramter @xmath99 , then an upper bound @xmath100 for @xmath99 in a dense subset of the paramter space implies the same upper bound for every @xmath99 .
we now consider the oscillations of vector - valued solutions of the system .
fix @xmath50 such that the affine line @xmath101 is not contained in @xmath39 .
then @xmath10 intersects @xmath39 in finitely many points .
let @xmath90 denote the complement of this intsection .
since the system is non - singular in @xmath90 , it admits an @xmath45-dimensional space @xmath102 of ( possibly multivalued ) vector - valued solution functions . to measure the oscillation of these solutions , we shall consider the number of intersections of a solution with an arbitrary fixed linear hyperplane .
formally , we define the linear space @xmath103 and the corresponding counting function @xmath104 when the system @xmath44 is clear from the context , we sometimes omit it from the notation and write @xmath105 .
we note that the counting function may in general be infinite
. we also remark that by triangulation , one may use to counting function @xmath97 to study the oscillation in more complicated domains . in this subsection
we introduce a class of systems of the form for which explicit bounds on the counting function @xmath106 may be derived .
this class constitutes the main object of study of the paper @xcite .
the system is said to be an @xmath107)-_q - system _ if @xmath44 is an @xmath37 matrix one - form defined over @xmath38 such that the following holds : 1 .
@xmath44 is integrable .
@xmath44 is regular .
@xmath44 is quasi - unipotent .
@xmath44 is defined over @xmath75 , has size @xmath108 , and coefficients of degree bounded by @xmath109 .
functions from the corresponding linear spaces @xmath110 are said to be _ q - functions_. the main interest in this class of systems stems from the following result of ( * ? ? ?
* theoren 8) , which plays the central role in the proof of theorem [ thm : bny ] .
[ thm : bny - qs ] let @xmath44 be an @xmath111-q - system .
then we have the following explicit bound , @xmath112 where @xmath113 .
we will also require a result concerning the order of a q - function near a singular point .
fix @xmath50 and let @xmath114 and @xmath115 a singular point of @xmath44 .
then , since @xmath44 is regular and quasiunipotent , @xmath116 admit an expansion @xmath117 , \mu\in\r.\ ] ] we call @xmath118 the _ order _ of @xmath88 at @xmath119 , and denote @xmath120 . if @xmath121 denotes a circular arc of radius @xmath70 and angle @xmath77 around @xmath122 , then @xmath123 the following proposition follows in a straightforward manner from the proof of theorem [ thm : bny - qs ] . [ prop : bny - order ] let @xmath44 be an @xmath111-q - system . fix some @xmath115 and let @xmath124
. then we have the following explicit bound , @xmath125 by it suffices to estimate the variation of argument of @xmath116 along @xmath121 ( in absolute value ) .
we list the appropriate references to @xcite .
the estimate follows immediately from principal lemma 33 and lemma 42 , noting the the normalized length of @xmath121 approaches @xmath126 as @xmath127 .
we remark that the bound of lemma 42 is stated for the variation of argument of @xmath88 , but it in fact applies to the absolute value of the variation of argument as well ( as is easily seen from the proof ) . in order to apply the theory of q - systems , and in particular theorem [ thm : bny - qs ] to the study of abelian integrals , it is necessary to produce a q - system that they satisfy .
the existence of such systems goes back to picard fuchs ( in the form ) , and to gauss
manin ( in the form ) .
explicit derivations of this system ( in the sense of subsection [ sec : complexity ] ) were given in @xcite .
for the convenience of the reader , we reproduce the relevant parts of the construction below . for proofs of all statements and further details see @xcite .
let @xmath128 denote the class of all hamiltonians of degree @xmath129 , @xmath130 where @xmath77 is a 2-multiindex .
then @xmath131 with @xmath132 provides an affine chart for @xmath128 .
let @xmath133 denote the affine curve defined by the equation @xmath134 .
for generic @xmath135 , the rank of the first homology group @xmath136 is @xmath137 .
one may choose a set of generators for this group over a fixed generic fibre @xmath138 , and transport them horizontally with respect to the gauss manin connection to obtain sections @xmath139 , ramified over a singular set @xmath140 .
under a further genericity assumption @xmath141 , we may assume further that the first cohomology group @xmath142 is generated by the monomial one - forms @xmath143 the _ period _ matrix @xmath144 is the @xmath37 matrix @xmath145 defined on @xmath146 and ramified over @xmath39 .
the period matrix satisfies a system of differential equations known as the picard fuchs system ( or gauss manin connection ) .
the following resut shows that this system is in fact a q - system .
[ thm : bny - gauss - manin ] the period matrix satisfies the equation @xmath147 , where @xmath44 is an @xmath111-q - system with @xmath148 let @xmath95 be the linear space spanned by @xmath149 ( possibly multivalued ) functions @xmath150 defined on a domain @xmath151
. denote by @xmath152 the space of polynomials of degree at most @xmath153 . by a slight abuse of notation
, we also denote by @xmath152 a @xmath154-q - system such that the entries of its fundamental solution matrix span the space @xmath152 ( such a system may easily be constructed ) .
the _ polynomial envelope _ of degree @xmath153 of the space @xmath95 is defined to be @xmath155 , \deg p_i{\leqslant}k.\ ] ] similarly , the _ polynomial envelope _ of a q - system @xmath44 is defined to be @xmath156 ( the tensor product of q - systems is discussed in section [ sec : transformations ] , transformation [ trans : tensor ] ) . to establish a link between the polynomial envelope and the study of abelian integrals we require the following result @xcite .
we use the notation of subsection [ sec : abelian - integrals ] . for a generic hamiltonian @xmath157 and for every polynomial one - form @xmath1 there exist univariate polynomials @xmath158 $ ] and bivariate polynomials @xmath159 $ ] such that @xmath160 where @xmath161 let @xmath162 denote the linear space of abelian integrals of forms of degree at most @xmath163 over the hamiltonian @xmath157 , and let @xmath164 denote the linear space of abelian integrals of the basic forms @xmath165 .
consider now an arbitrary polynomial one - form @xmath1 of degree at most @xmath163 .
let @xmath166 be a cycle on the @xmath108-level surface of @xmath157
. then @xmath167 and @xmath168 .
integrating over @xmath169 , @xmath170 [ cor : ai - envelope ] for a generic hamiltonian @xmath157 , @xmath171 in particular , at least when the hamiltonian is generic , @xmath172 is majorated by @xmath173 .
in this section we present the main result of the paper and deduce a corollary concerning the zeros of abelian integrals .
we begin by stating the general result of petrov - khovanskii .
our statement differs slightly from the usual formulation in order to facilitate the analogy to the uniform case . to simplify the notation , when speaking about an @xmath111-q - system we denote by @xmath99 the number of singular points of the system .
we record the following estimate , @xmath174 indeed , each singular point must be a pole of one of the @xmath175 entries of @xmath44 , and by degree considerations each entry may admit at most @xmath109 poles .
let @xmath176 be @xmath45 ( possibly multivalued and singular ) functions on @xmath177 , and let @xmath178 denote the linear space they span .
denote by @xmath179 the matrix @xmath180 suppose that @xmath181 is a rational matrix function of degree @xmath109 which is regular and quasiunipotent .
the following result can essentially be proved by combining the proofs of the petrov - khovanskii and the varchenko - khovanskii theorems ( see @xcite ) .
[ thm : pk - envelope ] under the conditions of the paragraph above , @xmath182 where @xmath183 is a constant depending only on @xmath184 ( for which a bound is not given ) .
in particular , the number of zeros of a function in the @xmath153-th polynomial envelope of @xmath178 grows at most linearly with @xmath153 .
the petrov - khovanskii result for abelian integrals , theorem [ thm : pk ] , follows from theorem [ thm : pk - envelope ] and corollary [ cor : ai - envelope ] for generic hamiltonians .
a slightly more refined argument is needed in order to remove the genericity assumption .
we exclude this argument from our presentation , as we shall soon see that our _ uniform _ version of the bound immediately extends from the generic case to the singular case .
we note that the system @xmath184 arising from the formulation of theorem [ thm : pk - envelope ] satisfies the various conditions required for a q - system , apart from the condition of being defined over @xmath75 .
this is not a coincidence .
in fact , the condition of being defined over @xmath75 is percisely the condition responsible for the emergence of _ uniform _ bounds in the class of q - systems .
we now state our main result .
[ thm : main ] let @xmath44 be an @xmath111-q - system . then @xmath185 note that , in contrast to theorem [ thm : pk - envelope ] , the bound in theorem [ thm : main ] is fully explicit . also note that while theorem [ thm : pk - envelope ] applies to a particular set of functions , theorem [ thm : main ] applies to families of functions depending ( as q - functions ) on an arbitrary number of parameters @xmath50 , and the bound is uniform over the entire family .
combining theorem [ thm : main ] with corollary [ cor : ai - envelope ] , we obtain an upper bound @xmath186 for the number of zeros of an abelian integral of degree @xmath163 over a generic hamiltonian @xmath157 of degree @xmath25 . by the semicontinuity of the counting function @xmath97 ( see remark [ rem : semicontinuity ] ) this bound extends over the entire class of hamiltonians , thus proving theorem [ thm : main - ai ] .
we note here that the implication above is a generally useful aspect of the theory of q - functions uniform bounds extend directly from the generic case to degenerate cases .
approaches based on compactness arguments usually require a more detailed analysis of the behavior near the singular strata ( see for instance the proof of theorem [ thm : pk - z ] in @xcite ) .
the approach employed by petrov and khovanskii in the proof of theorem [ thm : pk - envelope ] requires that we perform a number of transformations to the functions being considered .
our objective is to obtain uniform bounds by applying theorem [ thm : bny - qs ] .
it is therefore necessary to prove that the appropriate transformations can be carried it _ within _ the class of q - systems . in this section
we prove that this is indeed the case , and analyze the affect of each of the transformations on the parameters @xmath111 .
let @xmath44 denote an @xmath111-q - system , and let @xmath41 denote a fundamental solution for @xmath44 .
we assume that the base of the system is @xmath187 , with an affine chart @xmath188 .
[ trans : shift ] there exists an @xmath189-q - system @xmath190 defined over the base space @xmath191 , with affine chart @xmath192 , whose fundamental solution @xmath193 is given by @xmath194 and @xmath195 suppose that @xmath196 then @xmath197 since @xmath198 has an explicit solution @xmath193 , it is clear that @xmath198 is integrable .
it is also clear that the regularity and quasiunipotence of @xmath193 follows from that of @xmath41 . for the complexity analysis
, it remains only to notice that we increased the dimension of the base space by one , and that the complexity of the formula for @xmath198 is polynomial in the complexity and the maximal degree of the formula for @xmath44 , the dimension of @xmath44 and the dimension of the base space .
we remark that it is generally not possible to perform a shifting transformation by a specific _ fixed _
value @xmath199 . indeed ,
the formula for @xmath198 in this case would involve the specific value @xmath199 which may be irrational , while explicit algebraic formulas by our definitions may use only integer coefficients .
we circumvent this difficulty by extending the parameter space of the system with an additional parameter @xmath118 .
specific shifts of the system may be obtained by restricting @xmath118 to @xmath199 . the crucial condition which allows
this construction is that the system is not only a q - system for the fixed value @xmath199 , but rather it is a q - system with respect to the free parameter @xmath118 .
this technique is generally useful in the study of q - systems , and has already appeared in the context of the conformally invariant slope in @xcite .
we now consider the transformation of @xmath44 that corresponds to folding the @xmath49-plane by the transformation @xmath200 .
[ trans : fold ] there exists an @xmath189-q - system @xmath190 defined over the base space @xmath187 with affine chart @xmath201 , whose fundamental solution @xmath193 is given by @xmath202 where @xmath200 , and @xmath203 as in the proof of transformation [ trans : shift ] , it is clear that @xmath198 is integrable and regular . to prove quasi - unipotence ,
let @xmath54 be a small loop in the @xmath204 space .
if @xmath54 loops around a point with @xmath205 then it corresponds to a small loop in the @xmath206 plane , and the monodromy of @xmath207 around this loop is quasi - unipotent by the quasi - unipotence of @xmath44 .
if @xmath54 loops around a point with @xmath208 then @xmath209 corresponds to a small loop in the @xmath206 plane , and by the same reasoning we deduce that @xmath210 , the monodromy of @xmath211 along @xmath209 , is quasi - unipotent . but
@xmath212 , and a matrix whose square is quasi - unipotent is itself quasi - unipotent .
thus @xmath61 is quasi - unipotent as claimed . to explicitly define @xmath198 ,
suppose that @xmath196 then we may write @xmath213 since @xmath214 we may rewrite this expression in the form @xmath215 we now replace each occurence of @xmath216 by @xmath217 , giving an expression @xmath218 finally , since the second block in @xmath219 is equal to @xmath49 multiplied by the first block , we may rewrite this as @xmath220 which is an explicit expression for @xmath198 .
it is clear that the complexity of this expression is polynomial in @xmath221 , the base space dimension is unchanged , the dimension of @xmath198 is @xmath222 , and the maximal degree of the coefficients of @xmath198 is at most @xmath223 .
[ rem : fold - singularities ] if the singular points of @xmath44 for a specific value of @xmath50 form a set @xmath224 , then the singular values of @xmath198 form the set @xmath225 since @xmath226 and @xmath227 are the two critical values of the folding map .
we next consider symmetrization of @xmath44 around the real line .
this transformation was analyzed in @xcite .
we state here only the result and omit the proof ( which is straightforward ) . for convenience
we introduce the following notation .
the _ reflection _ of a function @xmath116 along the real line is given by @xmath228 if @xmath88 is multivalued then one may select an analytic germ of @xmath88 at some point on the real line , reflect this germ , and analytically continue the result . in cases where this choice is significant we shall state the point of reflection explicitly .
we will also use the @xmath229 notation for vector and matrix valued functions in the obvious way . in this paper
the reflection is always taken with respect to the time variable @xmath49 .
[ trans : symm ] there exists an @xmath189-q - system @xmath230 defined over the same base space as @xmath44 , whose fundamental solution @xmath193 is given by @xmath231 and @xmath232 the key feature of the symmetrization transform is that the corresponding solution spaces @xmath233 are closed under taking real and imaginary parts on the real line .
indeed , for any @xmath234 we have also @xmath235 , and therefore @xmath236 for completeness we also list the two canonical transformations of direct sum and tensor product . here
we let @xmath237 denote an @xmath238-q - system with fundamental solution @xmath239 for @xmath240 , defined over a common base space .
we again omit the proofs ( which are straightforward ) .
[ trans : sum ] there exists an @xmath189-q - system @xmath241 defined over the same base space as @xmath242 , whose fundamental solution is given by @xmath243 , and @xmath244 [ trans : tensor ] there exists an @xmath189-q - system @xmath245 defined over the same base space as @xmath242 , whose fundamental solution is given by @xmath246 , and @xmath247 here we use @xmath248 to denote the tensor product of @xmath242 as _ connections _ , but in order to avoid confusion we note that the matrix form representing this connection is in fact @xmath249 .
in this section we present the demonstration of theorem [ thm : main ] .
the proof follows the same strategy as the petrov - khovanskii proof of theorem [ thm : pk - envelope ] .
we first assume that all singular points of the system @xmath44 are real . in this case
it is possible to control the variation of argument by applying a clever inductive argument due to petrov . for the general case ,
we show that the system may be transformed to a system with real singular points , and invoke the preceding case .
[ prop : real - pk ] let @xmath44 be an @xmath111-q - system , and let @xmath50 be a paramter such that the singular locus of the system @xmath250 is contained in @xmath251 .
let @xmath252 and denote @xmath253 \\ \deg p_i(t ) { \leqslant}k \\
f_i\in l_{\l ' } \end{array } \right.\ ] ] finally , recall that we denote by @xmath99 the number of singular points of @xmath44 .
then @xmath254 let the domain @xmath90 and its boundary @xmath255 , partitioned as the union of the curves @xmath256 , be as indicated in figure [ fig : keyhole ] where the radius of each @xmath257 ( resp .
@xmath54 ) may be arbitrarily small ( resp .
notice that one segment of the real domain is in fact contained in @xmath90 ( indicated by a dotted line in the figure ) .
since any triangle avoiding the singular points can intersect at most one such segment , and since we can select @xmath90 to contain any single segment , it follows that to bound @xmath258 it will suffice to bound @xmath259 independently of the radii defining @xmath90 .
we proceed by induction on @xmath149 . for arbitrary @xmath149
, we proceed by applying the argument principle .
we first rewrite @xmath116 as @xmath264 where @xmath265 by theorem [ thm : bny - qs ] and the argument principle , @xmath266 we consider the variation of argument on each piece of @xmath255 separately .
the arcs @xmath267 are traversed in reverse orientation . therefore we need to bound the variation of argument along these arcs from below . by the contribution of each arc
approaches @xmath268 as @xmath127 . by proposition [ prop : bny - order ] , the order of each @xmath269 is bounded in absolute value by @xmath270 using we deduce that @xmath271 . therefore @xmath272 similarly , the arcs @xmath273 may be seen as small circular arcs around the point at infinity .
we argue as above , noting that in this case the order of each @xmath274 is bounded from below by @xmath275 .
it follows that @xmath276 .
therefore @xmath277 it remains to consider the variation of argument along the segments @xmath278 .
assume that @xmath279 is not purely real on @xmath278 ( otherwise there is no variation of argument ) .
the key observation is that @xmath280 where @xmath281 denotes the imaginary part taken with respect to the segment @xmath278 .
this fact , known as `` the petrov trick '' , is a simple topological consequence of the fact that the variation of argument of a curve contained in a half - plane is at most @xmath282 . using and noting that @xmath274 is real on the real line for every @xmath283
, we see that on @xmath278 @xmath284 where ( taking reflection with respect to @xmath278 ) , @xmath285 we used the fact that @xmath286 on @xmath278 .
let @xmath287 .
then @xmath198 is a @xmath288-q - system , and @xmath289 note that @xmath290 has the same singularities as @xmath250 , since the singular locus of @xmath250 is contained in @xmath251 , which is the set of fixed point for the reflection @xmath229 .
we may now apply the inductive hypothesis to @xmath291 , since the formula defining it only involves @xmath292 summands . @xmath293 using and summing up the variation of argument along @xmath255 using , and we finally obtain @xmath294 where all summands not involving @xmath153 are absorbed by the factor @xmath295 ( using the estimate ) .
in the proof above , we implicitly assume that @xmath116 does not vanish on the boundary of @xmath90 , so that the variation of argument is well defined .
this is a technical difficulty which can easily be avoided .
indeed , one can define the variation of argument by slightly deforming the boundary so that the zeros move to the exterior of @xmath90 , and taking the limit over the size of the deformation . with this notion ,
the estimates in the proof hold without any assumption .
every function @xmath297 may be written as @xmath298 \\
\deg p_j(t ) { \leqslant}k \\
f_j\in l_{\l ' } \end{array } \right.\ ] ] the right hand side is an expression of the form with @xmath299 . applying proposition [ prop : real - pk ] we obtain the bound stated above . to prove the general case
, we transform the system to have real singular points , and appeal to the result of the preceding subsection .
the transformation must be made within the class of q - systems , and uniform over the parameter space @xmath50 .
consider the following sequence of q - systems @xmath300 .
let @xmath301 , and define @xmath302 to be the system obtained from @xmath300 by applying the shifting transformation followed by the folding transformation ( we will denote the shifting parameter introduced at this step @xmath303 )
. set @xmath304 and @xmath305 .
we claim that for every @xmath50 , there is an appropriate choice of @xmath118 such that @xmath198 has real singularities for @xmath306 .
more specifically , we claim that for an appropriate choice of @xmath118 , the system @xmath300 will admit at most @xmath307 non - real singularities . to see this , we proceed by induction .
the original system @xmath44 admits at most @xmath99 singular points for any fixed value of the parameter @xmath50 . for step @xmath283 , select some non - real singular point @xmath108 of @xmath300 ( assuming there is such a point ) , and set @xmath308 .
then the shift transforms @xmath108 to a purely imaginary point .
the following fold transforms this point to the real line , transforms singularities already on the real line back to the real line , and only introduces new singularities at @xmath226 and @xmath227 ( see remark [ rem : fold - singularities ] ) . this concludes the induction .
a direct computations shows that @xmath198 is a @xmath309-q - system .
the number of singularities of the new system is at most @xmath310 .
it clearly suffices to prove that @xmath313 we may ignore the shift transform which ( for any fixed value of @xmath303 ) only introduces a constant additive factor to the time variable and does not affect .
henceforth we assume that @xmath302 is simply the fold of @xmath300 .
let @xmath49 denote the time variable of @xmath300 , and @xmath200 denote the time variable of @xmath302 .
for the sake of clarity we write @xmath314 to denote classes of polynomials in @xmath49 and @xmath217 respectively .
then @xmath315 where the last step follows directly from .
finally we observe that any triangular domain @xmath94 in the @xmath49-plane avoiding the singular locus of @xmath44 maps under the composed shifting and folding transforms to a domain covered by @xmath316 triangles in the time domain of @xmath190 .
this observation , combined with lemma [ lem : fold - envelope ] and corollary [ cor : main - real - singularities ] gives @xmath317 this concludes the proof of theorem [ thm : main ] .
the repeated - exponential nature of the bound in theorem [ thm : main - ai ] is clearly excessive .
we have therefore opted to emphasize clarity of exposition over optimality of the analysis .
in fact , a relatively straightforward ( though more technically involved ) computation using the proof of @xcite produces an improved estimate of only four repeated exponents . a key factor in the size of the bound is played by our construction ( following petrov and khovanskii ) of a composite folding transformation which moves all exisitng singularities of the system to the real line , while only introducing new singularities at real points . a more efficient construction of this type would yield better estimates .
we discuss a conjectural improvement of this type below .
let @xmath318 .
a polynomial @xmath319 is called a _ folding polynomial _ for @xmath320 if @xmath321 and @xmath319 admits only real critical values .
the change of variable @xmath322 , analogous to our basic folding transformation @xmath200 , moves the points of @xmath320 to the real line while only creating ramification points at the ( real ) critical values of @xmath319 .
the following conjecture , in this context , has already appeared in @xcite .
we note that the construction employed in the present paper , involving repeated shifting and squaring , produces folding polynomials of _ exponential _ degree . assuming the conjecture above , and generalizing our treatment of transformation [ trans : fold ]
, it is possible to improve our bound to a form involving only 3 repeated exponents . in any case
, the techniques of this paper rely heavily on the results of @xcite , and correspondingly the bounds obtained must be _ at least _ doubly - exponential .
it is very likely that this growth rate is still highly excessive .
furthr improvements will probably require completely new ideas .
sergei yakovenko .
quantitative theory of ordinary differential equations and the tangential hilbert 16th problem .
in _ on finiteness in differential equations and diophantine geometry _ , volume 24 of _ crm monogr .
_ , pages 41109 .
soc . , providence , ri , 2005 .
h. oadek .
, volume 67 of _ instytut matematyczny polskiej akademii nauk .
monografie matematyczne ( new series ) [ mathematics institute of the polish academy of sciences .
mathematical monographs ( new series)]_. birkhuser verlag , basel , 2006 . | an abelian integral is the integral over the level curves of a hamiltonian @xmath0 of an algebraic form @xmath1 .
the infinitesimal hilbert sixteenth problem calls for the study of the number of zeros of abelian integrals in terms of the degrees @xmath0 and @xmath1 .
petrov and khovanskii have shown that this number grows at most linearly with the degree of @xmath1 , but gave a purely existential bound .
binyamini , novikov and yakovenko have given an _
explicit _ bound growing doubly - exponentially with the degree .
we combine the techniques used in the proofs of these two results , to obtain an explicit bound on the number of zeros of abelian integrals growing linearly with @xmath2 . |
asymptotic analyses have been widely conducted in various research areas related to wireless communications .
although they do not quite provide the same information as complete ( non - asymptotic ) results , they usually give very useful insights , while being much more tractable and available for larger classes of models .
for example , the asymptotic coding gain in coding theory @xcite characterizes the difference of the signal - to - noise ratio ( snr ) levels between the uncoded system and coded system required to reach the same bit error rate ( ber ) in the high - snr regime ( or equivalently , the low - ber regime ) ; the diversity gain and the multiplexing gain introduced in @xcite are also high - snr asymptotic metrics that crisply capture the trade - off between the snr exponents of the error probability and the data rate in mimo channels ; in wireless networks , the asymptotic transmission capacity @xcite gives the network performance when the density of interferers goes to 0 , or equivalently , in the high signal - to - interference ratio ( sir ) regime .
these asymptotic analyses provide simple and useful results that capture important design trade - offs . in this paper , we focus on the asymptotic analyses of the sir distribution in wireless networks , which is a key metric that determines many other performance metrics , such as the achievable reliability , transmission rate , and the delay it is instrumental for the analysis and design of interference - limited wireless networks @xcite .
our analysis is not limited to one scenario but comprehensively covers a wide range of models , including both ad hoc and cellular networks , both singular and bounded path loss laws , and general stationary point processes and general fading unless otherwise specified . besides
, we also consider networks where each location of the transmitter ( or antenna ) has another transmitter ( or antenna ) colocated , which results in non - simple point processes . in all scenarios
, we mainly analyze the asymptotic properties of the sir distribution . for cellular networks with nakagami-@xmath0 fading and both singular and bounded path loss models , it has been observed in @xcite that the success probability , defined as the complementary cumulative distribution function ( ccdf ) of the sir , for different point processes are horizontally shifted versions of each other ( in db ) in the high - reliability regime .
generally , in non - poisson networks , the success probability is intractable . under this observation ,
however , we can obtain good approximations of the lower part of the sir distribution ( coverage probabilities above 3/4 ) for non - poisson networks if we know the result of the poisson networks and the corresponding shift amounts . for the tail of the sir distribution , a similar property holds , which has been proved in @xcite , if the singular path loss model is applied .
in general , the horizontal gaps of the sir distributions between a point process and the poisson point process ( ppp ) at both ends differ slightly , so for higher accuracy , the two asymptotic regimes should be treated separately .
this paper summarizes the known asymptotic properties , derives results for scenarios that have not been previously studied , and gives insight about the factors that mainly determine the behavior of the sir .
the reasons that we focus on the asymptotic sir analysis include the following : 1 .
it captures succinctly the performance of the various network models ( especially for the high - reliability regime ) .
it permits the isolation of the key network properties that affect the sir distribution .
3 . it gives insight into when it is safe to use the singular path loss model instead of a bounded one .
4 . it shows when a nearest - interferer approximation is accurate .
the tail determines whether the mean sir ( and higher moments ) exist . for poisson networks ,
the sir distribution has been derived in exact analytical form in a number of cases , namely for bipolar ad hoc networks with general fading @xcite , for ad hoc and cellular networks with successive interference cancellation @xcite , for multitier cellular networks ( hetnets ) with strongest - on - average base station association and base station cooperation without retransmissions @xcite and with retransmissions @xcite , for cellular networks with intra - cell diversity and/or base station silencing @xcite and for multitier cellular networks with instantaneously - strongest base station association @xcite . while for some specific assumptions ,
closed - form expressions are available , the results for the sir distribution typically involve one or more integrals . the only exact result for a non - poisson network is given in @xcite for cellular networks whose base stations form a ginibre process ; it contains several nested integrals and infinite sums and is very hard to evaluate numerically . from these exact results ,
simple asymptotic ones can often be derived , see , e.g. , @xcite , where results on the diversity gain are extracted from the more complicated complete results .
the true power of the asymptotic approach , however , becomes apparent when general non - poisson models are considered , for which essentially no exact results are available . in general ad hoc networks modeled using the bipolar model ,
the asymptotic sir distribution as the interferer density goes to 0 has been analyzed for rayleigh fading in @xcite and for general fading in @xcite . in (
* ch . 5 )
, the interference in general ad hoc networks has been analyzed .
relevant to our work here are the bounds on the ccdf of the interference and the asymptotic result on the interference distribution for both singular and bounded path loss models . in @xcite , we analyzed the asymptotic properties of the signal - to - interference - plus - noise ratio ( sinr ) distribution in general cellular networks in the high - reliability regime for nakagami-@xmath0 fading and both singular and bounded path loss models . in @xcite , a simple yet versatile analytical framework for approximating the sir distribution in the downlink of cellular systems
was proposed using the mean interference - to - signal ratio , while @xcite considers general cellular networks with general fading and singular path loss and studies the asymptotic behavior of the sir distribution both at 0 and at infinity .
the main contribution of the paper is a comprehensive analysis of the asymptotic properties of the ccdf of the sir @xmath1 , often referred to as the _ success probability _ or _
coverage probability_. regarding the transmitter / base station ( bs ) distributions , we do not restrict ourselves to simple point processes ( where there is only one point at one location almost surely ) , which are the ones used in almost all the literature , but also consider
duplicated-2-point " point processes .
the duplicated-2-point point processes are defined as the point processes where there are two points at the same location , i.e. , each point is duplicated .
the motivation to study this model is three - fold : ( 1 ) by comparing the results with those for standard ( simple ) network models , it becomes apparent whether the asymptotic behavior critically depends on the fact that the distances to the desired transmitter and to the interferers are all different ( a.s . ) .
( 2 ) there are situations where it is natural to consider a model where two nodes are at the same distance , such as when edge users ( who have two base stations at equal distance ) in cellular networks are analyzed or when spectrum sharing between different operators is assumed and the two operators share a base station tower .
( 3 ) the only kind of networks that are not captured by simple models are those that are formed by two fully correlated point processes .
the duplicated " models fills this gap .
the asymptotic properties of @xmath2 are summarized in table [ t_total ] with respect to @xmath3 as @xmath4 and @xmath5 for both singular and bounded path loss models , both ad hoc models and cellular models , and both simple point processes and duplicated-2-point point processes .
our results show that the asymptotic sir behavior is determined by two factors@xmath0 and @xmath6 .
@xmath0 is the nakagami fading parameter and @xmath7 , where @xmath8 is the path loss exponent .
as will be apparent from the proofs , the pre - constants are also known in some cases , not just the scaling behavior . the indicates that the results have been derived in the literature marked with the corresponding reference
; the marker ( * ) indicates that the results are only proven for the poisson case with rayleigh fading , while ( * * ) indicates that the results are only proven for the case of rayleigh fading and the duplicated-2-point point process where the distinct locations form a ppp .
.asymptotic properties ( simple " : simple point processes ; duplicated " : duplicated-2-point point processes ) [ cols="^,^,^",options="header " , ] [ t_sir0 ] we can simply modify the proof of theorem 1 in @xcite and prove that @xmath9 , as @xmath4 . here the nearest interferer at most at distance @xmath10 , since the point at @xmath10 is duplicated .
for the singular path loss model , we have @xmath11 + \mathbb{p}(r\geq b ) \mathbb{e}_{r}[\mathbb{p}\left ( h_0 < \theta h \right ) \mid r \geq b ] \nonumber\\ & = \mathbb{p}(r < b ) \mathbb{e}_{r}[\mathbb{p}\left ( h_0 < \theta h \ell(b)^{-1 } \ell(r)\mid r \right)\mid r < b ] + \mathbb{p}(r\geq b ) \mathbb{p}\left ( h_0 < \theta h \right ) .
\label{eq_adhoc}\end{aligned}\ ] ] from section [ sec::adsimpler0 ] , we know that the first term in is @xmath12 and the second term is @xmath13 .
so , @xmath14 , as @xmath4 . for the bounded path loss model
, we can apply the same methods as in the corresponding cases in section [ sec::adsimpler0 ] and obtain the results in table [ t_sir0 ] .
we have @xmath15 .
thus , @xmath16 . we can easily obtain the results in table [ t_sir0 ] .
we observe that in all cases , the lower tail is identical to the one when all interferers are considered .
for the duplicated point processes , this implies that duplicating only the nearest interferer ( as is the case for edge users in cellular networks ) again results in the same scaling . for both singular and bounded path loss models , we have @xmath17 since @xmath18
< \infty$ ] , for all @xmath19 , using the dominated convergence theorem and lhospital s rule , we obtain that @xmath20 , as @xmath5 . for the singular path loss model ,
we define @xmath21 as the distance from the origin to its nearest point of @xmath22 and assume @xmath23 .
we have @xmath24 \int_{\mathbb{r}^2 } \bar{f}_{h_0 } \left ( \|x\|^{\alpha } \right ) \mathrm{d}x , \quad \theta \to \infty \nonumber\\ & = \lambda \theta^{-\delta } \mathbb{e}\left[h^{-\delta } \right ]
\mathbb{e}\left[\bar{r}_2 ^ 2\right ] \pi \mathbb{e}(h^{\delta } ) , \quad \theta \to \infty , \end{aligned}\ ] ] where @xmath25 is a translated version of @xmath26 , @xmath27 follows by using the substitution @xmath28 , @xmath29 follows since @xmath30 , and @xmath31 follows by using the substitution @xmath32 .
thus , @xmath33 , as @xmath5 . for the bounded path loss model
, we have @xmath34 since @xmath35 < \infty$ ] and @xmath36 < \infty$ ] , for all @xmath19 , using the dominated convergence theorem and lhospital s rule , we obtain that @xmath20 , as @xmath5 . for both singular and bounded path loss models
, we can apply the same methods as in the corresponding cases in section [ sec::adsimpler0 ] and obtain the results in table [ t_sir0 ] .
we have @xmath15 .
thus , @xmath16 , and the results in table [ t_sir0 ] follow easily . since @xmath37 , we have @xmath38 , and thus @xmath39 .
consequently , @xmath4 , @xmath40 decays faster than or in the same order as @xmath41 ; also , since @xmath42 , as @xmath5 , @xmath43 decays slower than or in the same order as @xmath2 .
this is consistent with the results in tables [ t_total ] and [ t_sir0 ] . for non - trivial asymptotics , we need @xmath44 . in ad hoc models , for both the singular and bounded path loss models , whether @xmath45 is only determined by the distribution of the fading variable near @xmath46 . for the singular path loss model , whether @xmath47 is determined by both the tail of the fading distribution and the lower tail of the nearest - interferer distance distribution , while for the bounded path loss model , it is determined only by the tail of the fading distribution .
thus , for ad hoc models with bounded path loss , the decay order is only determined by the fading parameter @xmath0 . for simple ad hoc models with singular path loss
, we proved that the lower tail of the nearest - interferer distance distribution , other than the fading distribution , dominates the decay order and thus the decay order is only determined by @xmath6 . in cellular models , @xmath48 can be arbitrarily small and it always holds that @xmath49 .
for simple point processes , we proved that it is the fading variable , not the path loss exponent @xmath50 , that determines the decay order . for non - simple point processes , since @xmath15 , the fading variable alone determines the decay order . for non - trivial asymptotics
, we need @xmath51 . in simple ad hoc models , for both the singular and bounded path loss models , whether @xmath52 is determined only by the tail of the fading distribution and whether @xmath53 is determined by both the lower tail of the fading distribution and the tail of the nearest - interferer distance distribution .
we have proved that the fading distribution , rather than the nearest - interferer distance distribution , dominates the decay order .
the non - simple ad hoc models have the same asymptotics as the simple ones , since only the tail of the nearest - interferer distance distribution may affect the asymptotic property and in the non - simple case , the nearest - interferer distance can not be larger than @xmath10 . in simple cellular models , for the singular path loss model , whether @xmath52 is determined by the tail of the fading distribution and the lower tail of the distribution of @xmath48 , while @xmath53 is determined by both the lower tail of the fading distribution and the tail of the distribution of @xmath54 .
we have proved that the lower tail of the distribution of @xmath55 ( especially , that of @xmath48 ) , other than the fading distribution , dominates the decay order , so the decay order is a function of @xmath6 . for the bounded path loss model , whether @xmath52 is determined only by the tail of the fading distribution , while @xmath53 is determined by both the lower tail of the fading distribution and the tail of the distribution of @xmath54 .
we have proved that the fading distribution , rather than the distribution of @xmath54 , dominates the decay order . in non - simple cellular models , since @xmath15 ,
the fading distribution alone determines the decay order .
based on the insight obtained from the asymptotic results for the cases where the total interference is taken into account and where only the nearest interferer is considered , we can formulate three conjectures that succinctly summarize and generalize our findings for both cellular and ad hoc models assuming they hold .
let @xmath56 be the desired transmitter and @xmath57 the nearest interferer .
@xmath58 in words , this conjecture states that if the ( mean ) power of the nearest interferer can get arbitrarily large relative to that of the desired transmitter , then the outage probability only decays with @xmath59 as @xmath60 , and vice versa .
conversely , if the ratio is bounded away from @xmath46 , the only reason why the sir can get small is the fading , hence the fading statistics determine the scaling .
[ heavy tail ] @xmath61 in words , if the ( mean ) signal power can get arbitrarily large , the success probability has a heavy tail with exponent @xmath6 .
conversely , if the signal power is bounded , then the only reason why the sir can get big is the fading , hence the fading will ( co)determine the scaling ( and if only the nearest interferer is considered , the fading parameter @xmath0 alone determines the scaling ) . for rayleigh fading , @xmath62 in words ,
if the signal power is bounded and the fading is exponential , the tail will be exponential with @xmath63 .
we considered a comprehensive class of wireless network models where transmitters / bss follow general point processes and analyzed the asymptotic properties of the lower and upper tails of the sir distribution . in all cases ,
the lower tail of the sir distribution decays as a power law . only the path loss exponent or the fading parameter determines the asymptotic order for ad hoc networks ; only the fading parameter determines the asymptotic order for cellular networks .
this indicates that we can use the sir distribution of the ppp to approximate the sir distribution of non - poisson models in the high - reliability regime by applying a horizontal shift , which can be obtained using the mean interference - to - signal ratio ( misr ) defined in @xcite . for the tail of the sir distribution , only cellular network models with singular path loss have power - law tails . in these cases ,
we can approximate the tail of the sir distribution of non - poisson networks using the corresponding poisson result and the expected fading - to - interference ratio ( efir ) defined in @xcite . for cellular networks with bounded path loss and ad hoc networks , we mainly investigated the poisson case with rayleigh fading and showed that the tail of the sir distribution decays exponentially .
the exponential decay does not imply that it is not possible to approximate the tail of the sir distribution for non - poisson networks using the distribution of poisson networks .
we conjecture that the same approximation method as for polynomial tails can be applied , but a proof of the feasibility of this approach is beyond the scope of this paper and left for future work . moreover , we investigated the impact of the nearest interferer on the asymptotic properties of the sir distribution .
if the nearest interferer is the only interferer in the networks , we proved that the scaling of the lower tail of the sir distribution remains the same as when all interferers are taken into account , which means that the nearest interferer alone determines the decay order .
in contrast , for the tail of the sir distribution , the nearest interferer alone does not dominate the asymptotic trend , and the tail decays as a power law , with a light tail in most cases and a heavy tail for simple cellular networks with singular path loss .
these findings are summarized and generalized in three conjectures .
using the same method as in the proof of theorem 5.6 in @xcite , we can show that @xmath64 y^{-\delta } , \quad y \to \infty .
\label{heavy_tail_i_case2}\ ] ] the success probability can be rewritten as @xmath65 \nonumber\\ & \quad - \mathbb{e}_{h_0,h_1}\left [ \mathbb{p}\left ( \tilde{i } > h_0 b^{-\alpha } \theta^{-1 } - h_1 b^{-\alpha } \right ) \mathbf{1}(h_0 b^{-\alpha } \theta^{-1 } < h_1 b^{-\alpha } )
\mid h_0 , h_1 \right ] \nonumber\\ & = 1- \mathbb{e}_{h_0,h_1}\left [ \mathbb{p}\left ( \tilde{i } > h_0 b^{-\alpha } \theta^{-1 } - h_1 b^{-\alpha } \right ) \mathbf{1}(h_0 \theta^{-1 } \geq h_1 ) \mid h_0 , h_1 \right ] - \mathbb{p}\left ( h_0 \theta^{-1 } < h_1 \right ) .
\end{aligned}\ ] ] thus , @xmath66 + \mathbb{p}\left ( h_0 \theta^{-1 } <
h_1 \right ) } { \theta^{\delta } } \nonumber\\ & = \lim_{\theta \to 0 } \frac { \mathbb{e}_{h_0,h_1}\left [ \mathbb{p}\left ( \tilde{i } > h_0 b^{-\alpha } \theta^{-1 } - h_1 b^{-\alpha } \right ) \mathbf{1}(h_0 \theta^{-1 } \geq h_1 ) \mid h_0 , h_1 \right ] } { \theta^{\delta } } \nonumber\\ & \stackrel{(a)}{= } \mathbb{e}_{h_0,h_1}\left [ \lim_{\theta \to 0 } \frac { \mathbb{p}\left ( \tilde{i } >
h_0 b^{-\alpha } \theta^{-1 } - h_1 b^{-\alpha } \right ) \mathbf{1}(h_0 \theta^{-1 } \geq h_1 ) } { \theta^{\delta } } \mid h_0 , h_1 \right ] \nonumber\\ & \stackrel{(b)}{= } \mathbb{e}_{h_0,h_1}\left [ \lim_{\theta \to 0 } \frac{\pi \lambda b^2 \mathbb{e}[\left(h_a+h_b\right)^{\delta } ] } { 2 } \frac { \left ( h_0 \theta^{-1 } - h_1 \right)^{-\delta } \mathbf{1}(h_0 \theta^{-1 } \geq h_1 ) } { \theta^{\delta } } \mid h_0 , h_1 \right ] \nonumber\\ & = \frac{\pi \lambda b^2 \mathbb{e}[\left(h_a+h_b\right)^{\delta } ] } { 2 } \mathbb{e}_{h_0,h_1}\left [ \lim_{\theta \to 0 } \left ( h_0 - h_1 \theta \right)^{-\delta } \mathbf{1}(h_0 \geq h_1 \theta ) \mid h_0 , h_1 \right ] \nonumber\\ & = \frac { \pi \lambda b^2}{2 } \mathbb{e}[\left(h_a+h_b\right)^{\delta } ] \mathbb{e}[h^{-\delta } ] , \end{aligned}\ ] ] where @xmath27 follows from the dominated convergence theorem and @xmath29 follows from .
next , we show that nakagami-@xmath0 fading meets the fading constraints wherein @xmath67
. let @xmath68 .
we have @xmath69 = \int_0^{\infty } \frac{m^m}{\gamma(m ) } x^{m-1+\delta } e^{-mx } \mathrm{d}x = \frac{\gamma(m+\delta)}{m^{\delta}\gamma(m)},\end{aligned}\ ] ] @xmath70 = \int_0^{\infty } \frac{m^m}{\gamma(m ) } x^{m-1-\delta } e^{-mx } \mathrm{d}x = \frac{m^{\delta}\gamma(m-\delta)}{\gamma(m)},\end{aligned}\ ] ] and @xmath71 } { \theta^m } & = \lim_{\theta \to 0 } \frac{\mathbb{e}_{h_b}\left [ \int_0^{\theta h_b } \frac{m^m}{\gamma(m ) } x^{m-1 } e^{-mx } \mathrm{d}x \right ] } { \theta^m } \nonumber \\ & \stackrel{(a)}{= } \lim_{\theta \to 0 } \frac{\mathbb{e}_{h_b}\left [ \frac{m^m}{\gamma(m ) } \theta^{m-1 } h_b^{m } e^{-m\theta h_b } \right ] } { m \theta^{m-1 } } \nonumber \\ & = \frac{m^{m-1}}{\gamma(m ) } \lim_{\theta \to 0 } \mathbb{e}_{h_b}\left [ h_b^{m } e^{-m\theta h_b } \right ] \nonumber \\ & = \frac{m^{m-1}}{\gamma(m ) } \mathbb{e}_{h_b}\left [ h_b^{m } \right ] \nonumber \\ & = \frac{\gamma(2m)}{m \gamma(m)^2},\end{aligned}\ ] ] where @xmath27 follows by lhospital s rule .
so , @xmath73 } { \theta^{-(m+\delta ) } } = \lim_{\theta \to \infty } \frac { \frac{\partial } { \partial \theta } \mathbb{e } \left [ \bar{f}_{h}\left ( \theta \left(h_1 + r^{\alpha } \tilde{i } \right ) \right ) \right]}{- ( m + \delta ) \theta^{-(m+\delta+1 ) } } \nonumber \\ & = \lim_{\theta \to \infty}\mathbb{e } \left [ \frac { \frac{\partial } { \partial \theta } { f}_{h}\left ( \theta \left(h_1 + r^{\alpha } \tilde{i } \right ) \right ) } { ( m + \delta ) \theta^{-(m+\delta+1 ) } } \right ] \nonumber \\ & = \frac{1}{(m + \delta)\gamma(m ) } \lim_{\theta \to \infty}\mathbb{e } \left [ \theta^{2m+\delta } m^m \left(h_1 + r^{\alpha } \tilde{i } \right)^m e^{-\theta m \left(h_1 + r^{\alpha } \tilde{i } \right ) } \right ] \nonumber \\ & = \frac{m^m}{(m + \delta)\gamma(m ) } \lim_{\theta \to \infty}\mathbb{e } \left [ \theta^{2m+\delta } \left(\sum_{i = 0}^{m } \binom{m}{i } h_1^{i } \left ( r^{\alpha } \tilde{i}\right)^{m - i } \right ) e^{-\theta m \left(h_1 + r^{\alpha } \tilde{i } \right ) } \right ] \nonumber \\ & = \frac{m^m}{(m + \delta)\gamma(m ) } \sum_{i = 0}^{m } \binom{m}{i } \lim_{\theta \to \infty}\mathbb{e } \left [ \left ( \theta^{m+i } h_1^{i } e^{-\theta m h_1 } \right ) \left ( \theta^{m - i+\delta } \left ( r^{\alpha } \tilde{i}\right)^{m - i } e^{-\theta mr^{\alpha } \tilde{i } } \right ) \right ] \nonumber \\ & = \frac{m^m}{(m + \delta)\gamma(m ) } \sum_{i = 0}^{m } \binom{m}{i } \lim_{\theta \to \infty}\mathbb{e } \left [ \theta^{m+i } h_1^{i } e^{-\theta m h_1 } \right ] \mathbb{e } \left [ \theta^{m - i+\delta } \left ( r^{\alpha } \tilde{i}\right)^{m - i } e^{-\theta mr^{\alpha } \tilde{i } } \right ] \label{asymp_main}\end{aligned}\ ] ] where @xmath74 and @xmath75 are respectively the ccdf and the cdf of @xmath76 .
in the following , we evaluate @xmath77 $ ] and @xmath78 $ ] , separately . for @xmath77
$ ] , we have @xmath79 & = \lim_{\theta \to \infty } \int_0^{\infty } \frac{m^m}{\gamma(m ) } \left ( \theta^{m+i } x^{i } e^{-\theta m x } \right ) x^{m-1 } e^{-mx } \mathrm{d}x \nonumber\\ & = \lim_{\theta \to \infty } \frac{\gamma(m+i)}{m^i \gamma(m ) } \left(\frac{\theta}{1+\theta}\right)^{m+i } \int_0^{\infty } \frac{\left(m \left(1+\theta\right ) \right)^{m+i}}{\gamma(m+i ) } x^{m+i-1 } e^{-m(1+\theta)x } \mathrm{d}x \nonumber\\ & = \lim_{\theta \to \infty } \frac{\gamma(m+i)}{m^i \gamma(m ) } \left(\frac{\theta}{1+\theta}\right)^{m+i } \nonumber\\ & = \frac{\gamma(m+i)}{m^i \gamma(m)}. \label{asymp_fading}\end{aligned}\ ] ] when @xmath82 , let @xmath83 , and we have @xmath84 & = \lim_{\theta \to \infty } \frac{\delta \gamma(m - i)}{m^{m - i } } \frac { \frac{\partial } { \partial \theta } \mathbb{e } \left [ 1- \int_0^{\theta mr^{\alpha } \tilde{i } } \frac{1}{\gamma(m - i ) } x^{m - i -1 } e^{-x } \mathrm{d}x \right]}{\frac{\partial } { \partial \theta } \theta^{-\delta } } \nonumber\\ & \stackrel{(a)}{= } \lim_{\theta \to \infty } \frac{\delta \gamma(m - i)}{m^{m - i } } \frac { \mathbb{e } \left [ 1- \int_0^{\theta mr^{\alpha } \tilde{i } } \frac{1}{\gamma(m - i ) } x^{m - i -1 } e^{-x } \mathrm{d}x \right]}{\theta^{-\delta } } \nonumber\\ & = \lim_{\theta \to \infty } \frac{\delta \gamma(m - i)}{m^{m - i } } \frac { \mathbb{e } \left [ \bar{f}_{g_i}(\theta mr^{\alpha } \tilde{i } ) \right]}{\theta^{-\delta } } , \label{asymp_casei } \end{aligned}\ ] ] where @xmath27 follows by applying the lhospital s rule reversely and @xmath85 is the ccdf of @xmath86 .
using the same method as in the proof of theorem 4 in @xcite , we obtain @xmath87}{\theta^{-\delta } } \nonumber\\ & = \lim_{\theta \to \infty } \theta^{\delta } \sum_{x \in \phis } \bar{f}_{g_i}\left ( \theta m \|x\|^{\alpha } \left(\sum_{y \in \phis \setminus \{x\}}\left ( \left(h_{y,1}+h_{y,2}\right ) \|y\|^{-\alpha}\right)\right)\right ) \mathbf{1}(\phi(b(o,\|x\| ) = 0 ) ) \nonumber\\ & \stackrel{(b)}{= } \lim_{\theta \to \infty } \frac{\theta^{\delta } \lambda}{2 } \int_{\mathbb{r}^2 } \mathbb{e}_o^ ! \left [ \bar{f}_{g_i}\left ( \theta m \|x\|^{\alpha } \left(\sum_{y \in \phi_x}\left ( \left(h_{y,1}+h_{y,2}\right ) \|y\|^{-\alpha}\right)\right)\right ) \mathbf{1}\left(b(o,\|x\| ) \ ; { \rm empty } \right ) \right ] \mathrm{d}x \nonumber\\ & \stackrel{(c)}{= } \lim_{\theta \to \infty } \frac{\lambda}{2 } \int\limits_{\mathbb{r}^2 } \mathbb{e}_o^ ! \left [ \bar{f}_{g_i}\left ( m \|x\|^{\alpha } \left(\sum_{y \in \phi_{x\theta^{-\delta/2}}}\left ( \left(h_{y,1}+h_{y,2}\right ) \|y\|^{-\alpha}\right)\right)\right ) \mathbf{1}\left(b(o,\|x\|\theta^{-\delta/2 } ) \ ; { \rm empty } \right ) \right ] \mathrm{d}x \nonumber\\ & \stackrel{(d)}{= } \frac{\lambda}{2 } \int_{\mathbb{r}^2 } \mathbb{e}_o^ ! \left [ \bar{f}_{g_i}\left ( m \|x\|^{\alpha } i_{\infty } \right ) \right ] \mathrm{d}x \nonumber\\ & \stackrel{(e)}{= } \frac{\lambda}{2 } m^{-\delta } \mathbb{e}_o^ ! \left [ i_{\infty}^{-\delta } \right ] \int_{\mathbb{r}^2 } \bar{f}_{g_i}\left ( \|x\|^{\alpha } \right ) \mathrm{d}x \nonumber\\ & = \frac{\lambda}{2 } m^{-\delta } \mathbb{e}_o^ ! \left [ i_{\infty}^{-\delta } \right ] \pi \delta \int_0^{\infty } r^{\delta - 1 } \bar{f}_{g_i}\left ( r \right ) \mathrm{d}r \nonumber\\ & = \frac{\lambda}{2 } \pi m^{-\delta } \mathbb{e}_o^ ! \left [ i_{\infty}^{-\delta } \right ] \mathbb{e}[g_i^{\delta } ] \nonumber\\ & = \frac{\lambda \pi m^{-\delta } \gamma(m - i+\delta)}{2\gamma(m - i ) } \mathbb{e}_o^ ! \left [ i_{\infty}^{-\delta } \right ] , \label{asymp_-delta}\end{aligned}\ ] ] where @xmath29 follows from the campbell - mecke theorem , @xmath88 is a translated version of @xmath26 , @xmath31 follows by using the substitution @xmath28 , @xmath89 follows by the dominated convergence theorem and the fact that @xmath90 and thus @xmath91 , @xmath92 follows by using the substitution @xmath93 .
so , by substituting into , it yields that
@xmath84 = \frac{\delta \lambda \pi \gamma(m - i)}{2m^{m - i+\delta } } \mathbb{e}_o^ ! \left [ i_{\infty}^{-\delta } \right ] \mathbb{e}[g_i^{\delta } ] .
\label{asymp_casei_final}\end{aligned}\ ] ] when @xmath81 , let @xmath94 , and we have @xmath95 & = \lim_{\theta \to \infty } \mathbb{e } \left [ \theta^{\delta } e^{-\theta mr^{\alpha}\tilde{i } } \right ] \nonumber\\ & = \lim_{\theta \to \infty } \theta^{\delta } \mathbb{e } \left [ \bar{f}_{g_m}\left(\theta mr^{\alpha } \tilde{i } \right ) \right ] \nonumber\\ & \stackrel{(a)}{= } \frac{\lambda}{2 } \pi m^{-\delta } \mathbb{e}_o^ ! \left
[ i_{\infty}^{-\delta } \right ] \mathbb{e}[g_m^{\delta } ] \nonumber\\ & = \frac{\lambda}{2 } \pi m^{-\delta } \gamma(1+\delta ) \mathbb{e}_o^ ! \left [ i_{\infty}^{-\delta } \right ] , \label{asymp_case0}\end{aligned}\ ] ] where @xmath27 follows from .
r. k. ganti , j. g. andrews , and m. haenggi , `` high - sir transmission capacity of wireless networks with general fading and node distribution , '' _ ieee transactions on information theory _ ,
57 , no . 5 ,
31003116 , may 2011 .
m. haenggi , j. g. andrews , f. baccelli , o. dousse , and m. franceschetti , `` stochastic geometry and random graphs for the analysis and design of wireless networks , '' _ ieee journal on selected areas in communications _
27 , no . 7 , pp . 10291046 , sep .
a. guo and m. haenggi , `` asymptotic deployment gain : a simple approach to characterize the sinr distribution in general cellular networks , '' _ ieee transactions on communications _ , vol .
63 , no . 3 , pp .
962976 , mar . 2015 .
r. k. ganti and m. haenggi , `` asymptotics and approximation of the sir distribution in general cellular networks , '' _ ieee transactions on wireless communications _ , vol .
15 , no . 3 , pp .
21302143 , mar .
f. baccelli , b. blaszczyszyn , and p. mhlethaler , `` stochastic analysis of spatial and opportunistic aloha , '' _ ieee journal on selected areas in communications _ ,
27 , no . 7 , pp . 11051119 , sep
. 2009 .
x. zhang and m. haenggi , `` a stochastic geometry analysis of inter - cell interference coordination and intra - cell diversity , '' _ ieee transactions on wireless communications _ , vol . 13 , no . 12 , pp .
66556669 , dec . 2014 .
b. blaszczyszyn and h. p. keeler , `` studying the sinr process of the typical user in poisson networks by using its factorial moment measures , '' _ ieee transactions on information theory _ ,
61 , no . 12 , pp .
67746794 , dec . 2015 .
r. giacomelli , r. k. ganti , and m. haenggi , `` outage probability of general ad hoc networks in the high - reliability regime , '' _ ieee / acm transactions on networking _ , vol .
19 , no . 4 , pp . 11511163 ,
aug . 2011 .
a. guo , y. zhong , m. haenggi , and w. zhang , `` the gauss - poisson process for wireless networks and the benefits of cooperation , '' _ ieee transactions on communications _ ,
64 , no . 7 , pp . 29852998 , jul . | in the performance analyses of wireless networks , asymptotic quantities and properties often provide useful results and insights .
the asymptotic analyses become especially important when complete analytical expressions of the performance metrics of interest are not available , which is often the case if one departs from very specific modeling assumptions . in this paper , we consider the asymptotics of the sir distribution in general wireless network models , including ad hoc and cellular networks , simple and non - simple point processes , and singular and bounded path loss models , for which , in most cases , finding analytical expressions of the complete sir distribution seems hopeless . we show that the lower tails of the sir distributions decay polynomially with the order solely determined by the path loss exponent or the fading parameter , while the upper tails decay exponentially , with the exception of cellular networks with singular path loss . in addition
, we analyze the impact of the nearest interferer on the asymptotic properties of the sir distributions , and we formulate three crisp conjectures that if true
determine the asymptotic behavior in many cases based on the large - scale path loss properties of the desired signal and/or nearest interferer only .
stochastic geometry , point processes , asymptotics , interference , sir distribution . |
the recent observations of the cosmic microwave background ( cmb ) lead to a hubble constant of @xmath6 for the base six - parameter @xmath7cdm model ( planck collaboration 2013 , herafter p13 ) .
this value is in tension , at about the @xmath8 level , with the direct measurement of @xmath9 reported by r11 .
if these numbers are taken at face value , they suggest evidence for new physics at about the @xmath8 level ( for example , exotic physics in the neutrino or dark energy sectors as discussed in p13 ; see also wyman 2013 , hamann and hasenkamp 2013 ; battye and moss 2013 ; rest 2013 ; suyu 2013 )
. the exciting possibility of discovering new physics provides strong motivation to subject both the cmb and @xmath1 measurements to intense scrutiny .
direct astrophysical measurements of the hubble constant have a checkered history ( see , for example , the reviews by tammann , sandage and reindl 2008 ; freedman and madore 2010 ) .
the hubble space telescope ( hst ) key project led to a significant improvement in the control of systematic errors leading to ` final ' estimate of @xmath10 ( freedman 2001 ) . since then , two cepheid based programmes have been underway with the aim of reducing the error on @xmath1 : the _ supernovae and @xmath1 for the equation of state _ ( sh0es ) programme of r11 ( with earlier results reported in riess 2009 ) and the _ carnegie hubble program _ of freedman ( 2012 ) . in addition , other programmes are underway using geometrical methods , for example the megamaser cosmology project ( mcp ) ( reid 2013 ; braatz 2013 ) and the cosmological monitoring of gravitational lenses ( cosmograil ) project ( suyu 2010 , courbin 2011 ; trewes 2013 ) .
this paper presents a reanalysis of the r11 cepheid data .
the @xmath1 measurement from these data has the smallest error and has been used widely in combination with cmb measurements for cosmological parameter analysis ( _ e.g. _ hinshaw 2012 ; hou 2012 ; sievers 2013 ) .
the study reported here was motivated by certain aspects of the r11 analysis : the r11 outlier rejection algorithm ( which rejects a large fraction , @xmath11 , of the cepheids ) , the low reduced @xmath0 values of their fits , and the variations of some of the parameter values with different distance anchors , particularly the metallicity dependence of the period - luminosity relation .
the layout of this paper is as follows .
section 2 reviews the near - ir period - luminosity ( p - l ) relation of a sample of lmc cepheids . section 3 describes a reanalysis of the r11 sample using the maser distance to ngc 4258 as an anchor .
section 4 investigates the use of the lmc and mw cepheids as anchors .
section 5 investigates combinations of the three distance anchors and presents some internal consistency checks .
the conclusions are summarized in section 6 .
i will start with the lmc cepheids which i will use as a reference since the slope of the p - l relation is tightly constrained from this sample .
i use the 53 lmc cepheids with h - band magnitudes listed in persson ( 2004 ) and v , i magnitudes listed in sebo ( 2002 ) .
figure 1 shows the wesenheit magnitudes @xmath12 plotted against period @xmath13 ( in units of days ) .
the line shows a least - squares fit to @xmath14 _ i.e. _ minimising @xmath15 where @xmath16 is the error on @xmath17 and @xmath18 is the ` internal ' scatter that gives a reduced @xmath0 ( denoted @xmath19 in this paper ) of unity . since @xmath18 depends on the parameters @xmath20 and @xmath21 , the minimisation is performed iteratively until convergence .
the best fit parameters are @xmath22 the upper period limit is imposed because there is evidence that the p - l relation departs from a power law for periods @xmath23 days ( persson , 2004 ; freedman 2011 ; scowcroft 2011 ) . the lower period limit in ( [ lmc3b ] ) has been imposed because there have been some claims that the p - l relation at optical and near - ir wavelengths changes at periods less than 10 days ( ngeow 2009 ) .
the results of ( 4a ) and ( 4b ) and figure 1 show no evidence for any significant change in the power - law slope at low periods , in agreement with the results of persson ( 2004 ) .
evidently , a single power law is an extremely good fit to the lmc cepheids at least to periods of 60 days , and the slope of the p - l relation is determined to high accuracy .
as discussed by r11 , there are several reasons to expect outliers in the p - l relation . these include variables misidentified as classical cepheids , blended images , errors in crowding corrections and possible aliasing of the periods .
@xmath24 the h - band only p - l relations are fitted galaxy - by - galaxy to a power law with slope fixed at @xmath25 in the first iteration , weighted by the magnitude errors in table 2 of r11 .
cepheids with periods @xmath26 days are excluded .
once the outliers have been removed , r11 proceed to global fits using all of their galaxies , now adding a @xmath28 mag .
error in quadrature to the magnitude errors listed in their table 2 ( riess , private communication ) .
one of the consequences of adding this additional error term is that the @xmath19 values of the r11 global fits are always less than unity ( with typical values of @xmath29 ) so r11 rescale their covariance matrices by @xmath30 to compute errors on parameters . | i reanalyse the riess _
et al . _
( 2011 , hereafter r11 ) cepheid data using the revised geometric maser distance to ngc 4258 of humphreys _ et al . _
( 2013 , hereafter h13 ) .
i explore different outlier rejection criteria designed to give a reduced @xmath0 of unity and compare the results with the r11 rejection algorithm , which produces a reduced @xmath0 that is substantially less than unity and , in some cases , leads to underestimates of the errors on parameters .
i show that there are sub - luminous low metallicity cepheids in the r11 sample that skew the global fits of the period - luminosity relation .
this has a small but non - negligible impact on the global fits using ngc 4258 as a distance scale anchor , but adds a poorly constrained source of systematic error when using the large magellanic cloud ( lmc ) as an anchor .
i also show that the small milky way ( mw ) cepheid sample with accurate parallax measurements leads to a distance to ngc 4258 that is in tension with the maser distance .
i conclude that @xmath1 based on the ngc 4258 maser distance is @xmath2 , compatible within @xmath3 with the recent determination from for the base six - parameter cosmology .
if the h - band period - luminosity relation is assumed to be independent of metallicity and the three distance anchors are combined , i find @xmath4 , which differs by @xmath5 from the value .
the differences between the results and these estimates of @xmath1 are not large enough to provide compelling evidence for new physics at this stage . 0.15 truein * key words * : cosmology : distance scale , cosmological parameters
. 0.1 truein |
we shall investigate transport in networks with vertices and bonds which received a growing attention recently .
the networks of practical importance are those of nonlinear waveguides and and optical fibers @xcite , double helix of dna @xcite , josephson junction arrays with bose- einstein condensates @xcite , vein networks in leaves @xcite , etc .
major theoretical concern so far , however , is limited to solving stationary states of the linear schrdinger equation , and to obtaining the energy spectra in closed networks and transmission probabilities for open networks with semi - infinite leads @xcite .
only a few studies treats the nonlinear schrdinger equation on simple networks , which are still limited to the analysis of its stationary state@xcite . with introduction of the nonlinearity to the time - dependent schrdinger equation
, the network becomes to provide a nice playground where one can see interesting soliton propagations and nonlinear dynamics through the network @xcite , namely through an assembly of continuum line segments connected at vertices .
although there exist important analytical studies on the semi - infinite and finite chains @xcite , we find little exact analytical treatment of soliton propagation through networks within a framework of nonlinear schrdinger equation(nlse ) @xcite .
the subject is difficult due to the presence of vertices where the underlying chain should bifurcate or multi - furcate in general .
recently , with a suitable boundary condition at each vertex we developed an exact analytical treatment of soliton propagation through networks within a framework of nlse @xcite . under an appropriate relationship among values of nonlinearity at individual bonds , we found nonlinear dynamics of solitons with no reflection at the vertex .
we also showed that an infinite number of constants of motion are available for nlse on networks , namely the mapping of zakharov - shabat ( zs ) s scheme@xcite to networks was achieved .
the extension of the scenario to the discrete nlse ( dnlse ) is far from being obvious .
the standard dnlse is not integrable and the integrable variant of the continuum nonlinear schrdinger equation is the one proposed by ablowitz and ladik ( al)@xcite .
al equation is the appropriate choice for the zeroth order approximation in studying the soliton dynamics perturbatively in physically motivated models , such as an array of coupled optical waveguides @xcite and proton dynamics in hydrogen - bonded chains @xcite .
the dynamics of intrinsic localized modes in nonlinear lattices can be approximately described by al equation @xcite .
exciton systems with exchange and dipole - dipole interactions also reduce to al equation in some limiting cases@xcite .
the al chain is integrable by means of the inverse scattering transform , and , together with the toda lattice @xcite , constitutes a paradigm of the completely integrable lattice systems .
al equation for a field variable @xmath0 on one - dimensional ( 1-d ) chain is given by @xmath1 where @xmath2 is the strength of nonlinear inter - site interaction and @xmath3 denotes each lattice site on the chain .
this equation can be obtained from the canonical equation of motion with use of the non - standard poisson brackets .
equation ( [ al-0 ] ) has an infinite number of independent constants of motion and is completely integrable @xcite .
however , there is an ambiguity in generalizing the al model to networks : how can we define the inter - site interaction at each vertex in order to see the infinite - number of constants of motion in networks ? to keep the integrability of al equation , should any rule hold for the strength of nonlinearity on bonds joining at each vertex ? we shall resolve these questions in this paper and show how solitons of al equation on networks will be mapped to that of al equation on a 1-d chain .
once this mapping will be found , the integrability properties like the inverse scattering transform , bcklund transformation , etc , are automatically guaranteed , and will not be addressed in this paper . below we shall show the completely integrable case of the al equation on networks with strength of nonlinearity different from bond to bond . as a relevant issue , with use of reflectionless propagation of al soliton through networks , we shall evaluate the transmission probabilities on the outgoing bonds .
in section ii , using a primary star graph ( psg ) and defining a suitable equation of motion at the vertex , we shall address the norm and energy conservations . in section iii
, we shall show a basic idea of the soliton propagation along the branched chain , finding the connection formula at the vertex and the sum rule among the strengths of nonlinearity on the bonds , which guarantee the infinite number of constants of motions and complete integrability of the system under consideration . in section iv ,
the cases of generalized star graphs and tree graphs are investigated .
section v is devoted to the investigation of an injection of al soliton which bifurcates at the vertex and is decomposed into a pair of solitons with each propagating along the outgoing bonds , and we shall evaluate the transmission probabilities on the outgoing bonds . summary and discussions are given in section vi .
let us consider an elementary branched chain ( see fig.[star ] ) , namely , a primary star graph ( psg ) consisting of three semi - infinite bonds connected at the vertex @xmath4 .
and @xmath5 connected at a vertex @xmath4 . ]
we denote individual lattice sites as @xmath6 , where @xmath7 , and @xmath8 mean the bond s number and @xmath3 corresponds to a lattice site on each bond .
for the first bond @xmath9 , @xmath3 is numbered as @xmath10 , where @xmath11 means the branching point , i.e. , the vertex . for the second ( @xmath12 ) and third ( @xmath13 ) bonds
, @xmath3 varies as @xmath14 . @xmath15 and @xmath16 stands for the points nearest to the vertex . discrete nonlinear schrodinger equation ( dnlse )
_ la _ ablowitz - ladik ( al ) is defined on each bond except for the vicinity of the vertex as @xmath17 where @xmath18 .
it should be noted that @xmath19 may be different among bonds .
there is an ambiguity about the interaction around the vertex , which is resolved as follows : let s first introduce hamiltonian for psg as @xmath20 where at the virtual site @xmath21 we assume @xmath22 with appropriate coefficients @xmath23 and @xmath24 .
then eq.([eq1 ] ) can be obtained by the equation of motion @xmath25 at @xmath18 , with use of non - standard poisson brackets @xmath26 on the same footing as above , the equation of motions in eq.([net - al ] ) at ( 1,0 ) , ( 2,1 ) and ( 3,1 ) are given , respectively , as @xmath27 @xmath28 the solution is assumed to satisfy the following conditions at infinity : @xmath29 at @xmath30 and @xmath31 at @xmath32 for @xmath12 and 3 .
it is known that the norm conservation is one of the most important physical conditions in conservative systems . since eqs.([eq1]),([v10 ] ) and ( [ vk1 ] ) are available from hamilton s equation of motion with non - standard poisson brackets , the norm and energy conservations seem obvious .
below , however , we observe them explicitly .
extending the definition in the case of 1-d chain @xcite , the norm for psg is given as @xmath33 its time derivative is given by @xmath34 with @xmath35 for @xmath36 with use of eq .
( [ eq1 ] ) we have @xmath37 where @xmath38 implies a local current .
firstly one observes @xmath39 where @xmath40 means the summation over all sites on psg except for the points @xmath41 .
then , for @xmath42 , with use of eqs .
( [ v10 ] ) and ( [ vk1 ] ) we obtain @xmath43 and @xmath44 for @xmath45 . substituting eqs.([sum - up - akn ] ) , ( [ a10 ] ) and ( [ ak1 ] ) into eq.([dotn ] ) , we can see @xmath46 , i.e. , the norm conservation .
therefore , for any choice of values @xmath23 and @xmath24 the norm conservation turns out to hold well . on the other hand ,
the energy for psg is expressed in a symmetrical form as @xmath47.\end{aligned}\ ] ] to show that the energy is conservative , we see its time derivative @xmath48.\end{aligned}\ ] ] with use of eq .
( [ eq1 ] ) we have @xmath49\left(1+\gamma_1|\psi_{1,n}|^2\right)\nonumber\\ -&\psi_{1,-1}^*\dot\psi_{1,0},\end{aligned}\ ] ] and @xmath50\left(1+\gamma_1|\psi_{k , n}|^2\right)-\dot\psi_{k,1}^*\psi_{k,2}.\end{aligned}\ ] ] the first terms in the final expressions in eqs.([esum1 ] ) and ( [ esum23 ] ) are obviously pure - imaginary . substituting eqs .
( [ esum1 ] ) and ( [ esum23 ] ) into eq.([dote ] ) and using eqs.([v10 ] ) and ( [ vk1 ] ) , we find : @xmath51\nonumber\\ = 2{\rm re } & \left[\frac{1}{i}(1+\gamma_1|\psi_{1,0}|^2)\left(|\psi_{1,-1}|^2-|s_2\psi_{2,1}+s_3\psi_{3,1}|^2\right)\right.\nonumber\\ + \frac{1}{i}&\left.\sum_{k=1}^3(1+\gamma_k|\psi_{k,1}|^2)\left(s_k^2|\psi_{1,0}|^2-|\psi_{k,2}|^2\right)\right]= 0.\end{aligned}\ ] ] the last equality comes from the pure - imaginary nature of the expression in @xmath52 $ ] .
equation ( [ dote2 ] ) is nothing but the energy conservation .
thus we have proved that the norm and energy are conserved for any choice of values @xmath23 and @xmath24 . in general , however , other conservation rules do not hold . in the next sections we shall reveal a special case with appropriate choice of @xmath23 and @xmath24 which guarantees an infinite number of conservation laws .
among many possible choices of @xmath23 and @xmath24 , there is one special case in which an infinite number of constants of motion can be found and dnlse on psg becomes completely integrable . to investigate this case
, we shall first add to each bond @xmath53 ( @xmath54 ) a ghost - bond counterpart @xmath55 so that @xmath56 constitutes an ideal 1-d chain ( see fig .
[ ghost ] ) .
then we suppose that the soliton solution of al equation on psg is given by @xmath57 where @xmath58 are soliton solutions of dnlse with unit nonlinearity on the ideal 1-d chain ( @xcite ) : @xmath59 with @xmath3 being integers in @xmath60 .
the solutions of eq.([al on line ] ) may be different among three fictitious chains @xmath61 ( @xmath54 ) .
comparing eqs .
( [ v10 ] ) , ( [ vk1 ] ) and ( [ al on line ] ) , one can find at the vertex the following two equalities : @xmath62 @xmath63 noting the spatio - temporal behavior of soliton solutions and to guarantee the equality in eq.([comp - nvk1 ] ) , @xmath64 with @xmath45 should be satisfied for any time @xmath65 and for any integer @xmath3 , from which we obtain @xmath66 and @xmath67 namely , the solution @xmath58 should be bond - independent . with use of eqs.([s - k ] ) and
( [ q - kn ] ) in eq .
( [ comp - nv10 ] ) we have the sum rule among nonlinearity coefficients @xmath68 , @xmath69 and @xmath70 : @xmath71 equations ( [ s - k ] ) , ( [ q - kn ] ) and ( [ sumrule - g ] ) are the necessary and sufficient conditions to see eqs.([comp - nv10 ] ) and ( [ comp - nvk1 ] ) .
thus , under the sum rule for nonlinearity coefficients in eq.([sumrule - g ] ) , the solution on psg is given by a common ( bond - independent ) soliton solution of eq.([al on line ] ) multiplied by square root of the inverse nonlinearity coefficient .
for example , the soliton incoming through the bond @xmath72 is expected to smoothly bifurcate at the vertex and propagate through the bonds @xmath73 and @xmath5 , as we shall see in fig .
[ numerical ] . in the case
that @xmath68 , @xmath69 and @xmath70 break the sum rule , we shall see a completely different nonlinear dynamics of solitons such as their reflection and emergence of radiation at the vertex , as will be shown in fig.[numerical - ex ] .
the initial value problem for such a case is outside the scope of the present work .
we also note that the parameters @xmath23 and @xmath24 would correspond to @xmath74 and @xmath75 , respectively , in the preceding work @xcite , although the derivations of the connection formula at the vertex are quite different between the continuum and discrete systems . in fact ,
@xmath23 and @xmath24 are introduced to define the inter - site interaction at the vertex and are not obtained from the norm and energy conservations , in contrast to the case of networks consisting of continuum segments@xcite .
it is well known that ablowitz - ladik ( al ) equation on the 1-d chain has an infinite number of constants of motion .
now we shall proceed to obtain an infinite number of constants of motion for general solutions of al equation on psg .
first of all , it should be noted that the solution on psg can now be written as @xmath76 where @xmath77 stands for a general solution of al equation ( [ al on line ] ) and is restricted to each bonds @xmath53 ( @xmath54 ) . while we already proved the conservation of energy , we can generalize it to the general case : without taking the complex conjugate , eq .
( [ net - haml ] ) can be explicitly written as @xmath78 substituting eq.([gensol ] ) into eq.([gl - energy1 ] ) , @xmath79 is rewritten as @xmath80 using the value @xmath81 in eq.([s - k ] ) and the sum rule in eq.([sumrule - g ] ) , eq.([gl - energy2 ] ) reduces to the constant for the ideal 1-d chain @xcite : @xmath82 therefore @xmath79 in eq.([gl - energy1 ] ) is a constant of motion , and its real and imaginary parts imply the energy and current , respectively .
for other higher - order conservation rules , we can write them as @xmath83 with @xmath84 defined as expansion coefficients of the expression ( see ablowitz & ladik @xcite ) @xmath85 where @xmath86 are given by @xmath87 @xmath88 the relations ( [ cl3 ] ) and ( [ cl4 ] ) are obtained by solving eq .
( 4.15 ) in @xcite , i.e. , @xmath89 recursively with use of the expansion @xmath90 the right - hand side of eq .
( [ conslaws ] ) includes some undefined field variables in the ghost bond regions which must be defined as @xmath91 the conservation laws in eq .
( [ conslaws ] ) follows from the nature of solutions ( [ gensol ] ) and the sum rule for nonlinearity coefficients ( [ sumrule - g ] ) . for @xmath92
we obtain current and energy conservation laws . at @xmath93
we obtain higher order conservation laws .
some of higher - order constants of motion are as follows : @xmath94 @xmath95,\end{aligned}\ ] ] where field variables at lattice sites of the ghost bonds are defined in eq .
( [ undef ] ) .
now we proceed to explore soliton solutions of dnlse on other types of graphs and explore the sum rule and conservation rules for solitons to propagate through these graphs .
the above treatment on psg is also true for more general star graphs consisting of @xmath96 semi - infinite bonds connected at a single vertex . in such cases , the initial soliton at an incoming bond @xmath72 splits into @xmath97 solitons in the remaining bonds , and the extended version of eq .
( [ sumrule - g ] ) is @xmath98 the solution is given by the equations @xmath99 where @xmath100 for the first bond ( @xmath9 ) and @xmath101 for other bonds ( @xmath102 ) .
@xmath103 is a soliton solution of eq .
( [ al on line ] ) .
conservation laws for this graph can be obtained analogously as in the case of psg . , and @xmath104 with @xmath105 . ] another example of the graph for which the soliton solution of dnlse can be obtained analytically is a tree graph in fig .
[ tree-0 ] .
now we shall provide a soliton solution in this case .
we denote bonds of graph as @xmath106 and number the lattice sites on this bonds as @xmath107 . on each branching point
we assume the following conditions are hold @xmath108 the solution is given by @xmath109 here @xmath110 number of lattice sites that soliton pass through from @xmath72 to @xmath111 . for the tree graph it is defined as @xmath112 below ,
applying the induction method we give a proof of conservation laws for soliton solutions of al on tree graph .
let us denote the tree graph as @xmath113 and assume the conservation laws to hold in @xmath113 : @xmath114 .
then we construct an enlarged tree graph in the following way : first we choose the arbitrary point @xmath115 in the one of right - most semi - infinite chain @xmath116 as a new branching point .
cut of semi - infinite part of this bond at the point @xmath115 and attach @xmath117 semi - infinite bonds to this point .
namely the bond @xmath116 is now replaced by a finite bond @xmath118 connected with @xmath117 semi - infinite bonds @xmath119 , with @xmath120 . for the enlarged tree graph ,
constants of motion are given by @xmath121 it is clear that the final expression becomes constant under the sum rule ( [ sumrule ] ) .
thus , starting from psg in fig .
[ star ] and repeating the above procedure , we can get the conservation rule for all tree graphs .
a relevant issue of the above discoveries is the transmission probability against injection of a single soliton .
here we calculate transmission probabilities for a single soliton which is incoming through a semi - infinite bond @xmath72 and outgoing through the other semi - infinite bonds @xmath122 . a single ( bright ) soliton on a graph , which takes the general form as in eqs . ( [ gensol ] ) , ( [ sol - gen - star ] ) and ( [ treesoliton ] ) ,
is described with use of al soliton with @xmath123 @xcite : @xmath124 lying on individual bonds @xmath125 is given by @xmath126 \nonumber\\ & \times e^{-i(\omega t+\alpha n+\phi_{0 } ) } , \nonumber\\ & \qquad n\in b_l , \
l=1,2,3 , \cdots , n,\end{aligned}\ ] ] where @xmath127 , @xmath128 , @xmath129 , @xmath130 , @xmath131 and @xmath132 are bond - independent parameters characterizing frequency , velocity , wave number , inverse width of the soliton , initial phase and initial center of mass , respectively . equation ( [ sol ] ) indicates that a narrow soliton travels faster than wider ones with the same @xmath133 .
it should be noted that parameter values are common to each bond , except for @xmath134 .
choosing the simplest network psg in fig.[star ] , we shall give conservative quantities for the solution in eq .
( [ sol ] ) under the sum rule in eq.([sumrule - g ] ) .
first of all , the norm in eq.([net - norm ] ) turned out to be reduced to the one for the 1-d chain with the nonlinearity constant @xmath68 and thereby is given by @xmath135 equation ( [ norm - sol ] ) indicates that a narrow soliton has a larger norm than wider ones . as for the energy @xmath136 and current @xmath137 , it is convenient to evaluate the combined quantity @xmath79 in eq .
( [ gl - energy1 ] ) with use of @xmath23 and @xmath24 given by eq .
( [ s - k ] ) .
in fact we have @xmath138 substituting eq .
( [ sol ] ) into eq .
( [ gl - energy1 ] ) and using the sum rule in eq .
( [ sumrule - g ] ) , one obtains @xmath139 and @xmath140 as is seen from eq .
( [ sol ] ) , the center of mass of the soliton ( cms ) on each bond @xmath125 is located at @xmath141 at @xmath142 .
however , lattice points on the individual semi - infinite bonds are defined on the limited interval . in particular , on outgoing bonds @xmath122 , their lattice points @xmath3 are defined in the interval @xmath143 .
if @xmath144 , therefore , cms on @xmath122 is initially located outside of the real bonds . in such cases
we call the soliton as a `` ghost soliton '' .
when cms belongs to a real bond we use a term `` real soliton '' . in fig . [ ghost ] which corresponds to psg in fig.[star ] ,
ghost solitons are plotted by broken curve while real ones by solid line .
the soliton dynamics here is governed by a single characteristic time @xmath145 .
while for @xmath146 the soliton at @xmath72 is a real one and those at @xmath73 and @xmath5 are ghosts , for @xmath147 the soliton at @xmath72 is a ghost and those at @xmath73 and @xmath5 are real . at @xmath142 with @xmath148 ,
the soliton lying on the bond @xmath72 is exclusively responsible for the norm @xmath96 . on the other hand , at @xmath149 ,
the solitons running through the bonds @xmath73 and @xmath5 are exclusively responsible for the norm
. therefore we can naturally define transmission probabilities at @xmath150 .
satisfying the sum rule in eq.([sumrule - g ] ) .
space distribution of wave function probability is depicted in every time interval @xmath151 with time used commonly in branches 2 and 3 .
abscissa represents discrete lattice coordinates defined in fig.[star ] .
initial profile is ablowitz - ladik soliton in eq.([sol ] ) at @xmath142 with parameters @xmath152 .
time difference in numerical iteration is @xmath153 bottom panel shows the time dependence of partial norms at each of 3 branches . ]
satisfying the sum rule in eq.([sumrule - g ] ) .
space distribution of wave function probability is depicted in every time interval @xmath151 with time used commonly in branches 2 and 3 .
abscissa represents discrete lattice coordinates defined in fig.[star ] .
initial profile is ablowitz - ladik soliton in eq.([sol ] ) at @xmath142 with parameters @xmath152 .
time difference in numerical iteration is @xmath153 bottom panel shows the time dependence of partial norms at each of 3 branches . ,
width=188 ] in psg .
symbols and lines denote numerical and theoretical results , respectively . a solid line with @xmath154 and a broken line with @xmath155 correspond to @xmath156 and @xmath157 , respectively . ] in general networks , transmission probability for an arbitrary sem - infinite bond @xmath158 at discrete time @xmath159 that makes @xmath160 integers are defined as @xmath161 at @xmath162 , @xmath163 on the last line in eq.([dtp_def ] ) tends to @xmath164 and this summation gives @xmath96 , i.e. , the normalization of the soliton in the ideal 1-d chain with the nonlinearity coefficient @xmath68
. therefore @xmath165 under the sum rules as in eqs .
( [ sumrule - g ] ) , ( [ beta - g ] ) and ( [ sumrule ] ) we have the unitarity condition @xmath166 where the summation is taken over the semi - infinite bonds except for @xmath72 .
the result in eq.([dtp2 ] ) means that the transmission probability is inversely proportional to the strength of nonlinearity in outgoing semi - infinite bonds .
we have checked this result using a numerical simulation of the discrete nonlinear schrdinger equation(dnlse ) on psg in fig.[star ] : we numerically iterated eqs.([eq1 ] ) , ( [ v10 ] ) and ( [ vk1 ] ) with use of eq.([s - k ] ) and chose the initial profile in eq.([sol ] ) with @xmath68 and @xmath167 as an incoming soliton .
figure [ numerical ] shows the result in the case that the sum rule in eq.([sumrule - g ] ) is satisfied : the soliton starting at lattice point @xmath168 in the branch 1 enters the vertex at @xmath169 and is smoothly split into a pair of smaller solitons in the branches 2 and 3 with no reflection at the vertex .
the velocity and width of the soliton have the definite value common to all bonds , and the squared peak value of the soliton is proportional to @xmath19 , which are consistent with the result in eq.([sol ] ) .
bottom panel in fig.[numerical ] shows the time dependence of partial norms at each of 3 branches .
with increasing time , the partial norms at branches 2 and 3 converge to the transmission probabilities in eq.([dtp2 ] ) . in fig.[transm ]
transmission probabilities @xmath156 and @xmath157 are plotted as a function of @xmath170 in the wider range of @xmath68 and @xmath68 in the case satisfying the sum rule in eq.([sumrule - g ] ) .
we can confirm the linear law predicted in eq.([dtp2 ] ) . , which breaks the sum rule .
initial profile and parameter values are the same as in fig.[numerical ] . in top panel ,
broken curves indicate a propagation of the reflected soliton .
bottom panel shows the time dependence of partial norms at each of 3 branches . ] , which breaks the sum rule .
initial profile and parameter values are the same as in fig.[numerical ] . in top panel ,
broken curves indicate a propagation of the reflected soliton .
bottom panel shows the time dependence of partial norms at each of 3 branches.,width=188 ] figure [ numerical - ex ] shows the result in the case that the sum rule is broken : @xmath171 . in this case
the soliton starting at lattice point @xmath168 in the branch 1 enters the vertex at @xmath169 , but is accompanied with both reflection and emergence of radiation at the vertex .
it is very interesting that the velocity of the self - organized soliton have the definite value common to all bonds .
in particular , the reflected soliton at the branch 1 has the same magnitude of velocity as that of the incident soliton .
with increasing time , the partial norms at branches 1 , 2 and 3 would converge to the reflection ( on @xmath72 ) and transmission probabilities ( on @xmath73 and @xmath5 ) .
for some other choice of @xmath68,@xmath69 and @xmath70 that breaks the sum rule ( which is not shown here ) , the asymptotically(@xmath149)-equal velocity of solitons running on all three semi - infinite bonds can also be observed and provides an open question to be resolved in due course .
we have derived conditions under which ablowitz - ladik ( al ) type discrete nonlinear schrdinger equation ( dnlse ) on simple networks is mapped to the original one on the ideal 1-d chain and becomes completely integrable . here
the strength of cubic nonlinearity is different from bond to bond , and networks are assumed to have at least two semi - infinite bonds with one of them used as an incoming bond .
our findings are : ( 1 ) the solution on each bond is a part of the universal ( bond - independent ) soliton solution of the completely - integrable dnlse on the 1-d chain , but is multiplied by the inverse of square root of bond - dependent nonlinearity ; ( 2 ) the inverse nonlinearity at an incoming bond should be equal to the sum of inverse nonlinearities at the remaining outgoing bonds ; ( 3 ) with use of the above two findings , there exist an infinite number of constants of motion . the parameters @xmath23 and @xmath24 , which played an essential role in deriving the connection formula , are introduced to define the inter - site interaction at the vertex and are not obtained from the norm and energy conservations , in marked contrast to the case of networks consisting of continuum segments@xcite .
the argument on a branched chain or a primary star graph ( psg ) is generalized to general star graphs and tree graphs by using the induction method . as a practical issue , with use of al soliton injected through the incoming bond , we obtain transmission probabilities inversely proportional to the strength of nonlinearity on the outgoing bonds .
kivshar and g.p .
agarwal , _ optical solitons : from fibers to photonic crystals _ ( academic press , san diego , 2003 ) . l.v .
yakushevich , a.v .
savin and l.i .
manevitch phys .
rev . e*66 * , 016614 ( 2002 ) .
f. s. cataliotti , s. burger , c. fort , p. maddaloni , f. minardi , a. trombettoni , a. smerzi and m. inguscio , science * 293 * , 843 ( 2001 ) .
f. corson , phys .
* 104 * , 048703 ( 2010 ) .
e. katifori_et al . , _ phys .
lett . * 104 * , 048704 ( 2010 ) .
p. exner and p. eba , phys
128a * , 493 ( 1988 ) .
p. exner and p. eba , rep .
* 28 * , 7 ( 1989 ) .
t. kottos and u. smilansky , ann . phys .
( ny ) * 274 * , 76 ( 1999 ) ; s. gnutzmann and u. smilansky , adv . in phys . *
55 * , 527 ( 2006 ) .
p. kurasov and f. stenberg , j. phys .
a*35 * , 101 ( 2002 ) ; p. kurasov and m. nowaczyk , j. phys .
a * 38 * , 4901 ( 2005 ) .
p. kuchment and o. post , comm .
* 275 * , 805 ( 2007 ) .
p. exner , j.p .
keating , p. kuchment and t. sunada , _ analysis on graphs and its applications _
soc . , 2008 ) s. gnutzmann , u. smilansky and s. derevyanko , phys .
a * 83 * , 033831 ( 2011 ) .
r. adami , c. cacciapuoti , d. finco and d. noja , arxiv:1104.3839v1[math - ph ] ( 2011 ) .
s. flach and c. r. willis , phys . rep .
* 295 * , 181 ( 1998 ) .
r. burioni , d. cassi , p. sodano , a. trombettoni and a. vezzani , chaos * 15 * , 043501 ( 2005 ) ; physica d * 216 * , 71 ( 2006 ) .
a.s . pikovsky and d. l. shepelyansky , phys .
* 100 * , 094101 ( 2008 ) .
g. kopidakis , s. komineas , s. flach and s. aubry , phys . rev . lett .
* 100 * , 084103 ( 2008 ) .
m.j.ablowitz and h.segur , j. math .
* 16 * , 1054 ( 1975 ) .
a.s.fokas , a.r.its and l - y sung , nonlinearity * 18*,1771 ( 2005 ) . j.i.ramos and f.r.villatoro , mathl .
. modelling . * 20 * , 31 ( 1994 ) . a.s.fokas and a.r.its , j. phys .
a : math . gen . * 37 * , 6091 ( 2004 ) . c. sulem and p.l .
sulem , _ the nonlinear schrdinger equation _ ( springer , new york , 1999 )
. m. j. ablowitz , b. prinari and a. d. trubatch , _ discrete and continuous nonlinear schrdinger systems _
( university press , cambridge , 2004 ) . v.b .
zakharov and a.b .
shabat , sov .
* 34 * 62 ( 1972 ) .
z. sobirov , d. matrasulov , k. sabirov , s. sawada and k. nakamura , phys .
e * 81 * , 066602 ( 2010 ) .
m. j. ablowitz and j. f. ladik , j. math .
* 16 * , 598 ( 1975 ) .
m. j. ablowitz and j. f. ladik , j. math .
* 17 * , 1011 ( 1976 ) .
g. biondini and g , hwang , inverse probl .
* 24 * , 065011 ( 2008 ) .
a. b. aceves , c. de angelis , t. peschel , r. muschall , f. lederer , s. trillo and s. wabnitz , , phys .
e * 53 * , 1172 ( 1996 ) .
s. yomosa , j. phys .
jpn . * 52 * , 1866 ( 1983 ) .
s. kivshar and b.a .
malomed , rev .
phys . * 63 * , 761 ( 1989 ) .
s. takeno , j. phys .
japan * 61 * , 1433 ( 1992 ) .
konotop and s. takeno , phys .
b * 55 * , 11342 ( 1997 ) .
m. toda , phys . rep . * 18 * , 1(1973 ) ; _ theory of nonlinear lattices _ ( springer , berlin , 1981 ) . | we elucidate the case in which the ablowitz - ladik ( al ) type discrete nonlinear schrdinger equation ( nlse ) on simple networks ( e.g. , star graphs and tree graphs ) becomes completely integrable just as in the case of a simple 1-dimensional ( 1-d ) discrete chain . the strength of cubic nonlinearity is different from bond to bond , and networks are assumed to have at least two semi - infinite bonds with one of them working as an incoming bond .
the present work is a nontrivial extension of our preceding one ( sobirov _ et al _ , phys .
rev .
e * 81 * , 066602 ( 2010 ) ) on the continuum nlse to the discrete case .
we find : ( 1 ) the solution on each bond is a part of the universal ( bond - independent ) al soliton solution on the 1-d discrete chain , but is multiplied by the inverse of square root of bond - dependent nonlinearity ; ( 2 ) nonlinearities at individual bonds around each vertex must satisfy a sum rule ; ( 3 ) under findings ( 1 ) and ( 2 ) , there exist an infinite number of constants of motion . as a practical issue , with use of al soliton injected through the incoming bond , we obtain transmission probabilities inversely proportional to the strength of nonlinearity on the outgoing bonds . |
since its advent , density - functional theory ( dft ) @xcite is routinely applied for calculating the electronic , magnetic , spectroscopic and thermodynamic properties of atoms , molecules and materials in ground and excited states . in the last couple of decades ,
studying excited- states employing dft has become the main research interest @xcite .
thus one of the most natural approach to do excited - state dft is to adopt the time - independent density functional formalism @xcite in which the individual excited - state energies are determined from the stationary states of the energy density functional .
however , the question is whether there exists any such functional(s ) for excited states analogous to the ground - state . not only energy functionals but also the most fundamental and essential requirement for excited - state density functional theory ( @xmath0dft ) is to establish the one - to - one mapping similar to the hohenberg - kohn theorem which is the main intent of the present work . although the issue of density @xmath1 to potential @xmath2 mapping for excited states has been addressed in the past @xcite , but the question still remains unanswered .
so the current work will answer the critiques of density - to - potential mapping based on the generalized / unified constrained search(cs ) due to perdew - levy(pl ) @xcite , grling @xcite , levy - nagy(ln ) @xcite , samal - harbola(sh ) @xcite and ayers - levy - nagy @xcite . in the present work , we will critically analyse and make furtherance to the @xmath0dft ideas proposed by samal and harbola @xcite . according to it , ( i ) the cs approach can be extended to excited - state in the light of the stationary state formalism of grling @xcite and variational @xmath0dft formalism by levy - nagy @xcite ; ( ii ) within the variational @xmath0dft formalism , the construction of the kohn - sham(ks ) system by comparing only the ground - state density is insufficient and ca nt explain the existence of multiple potentials ; ( iii ) the density - to - potential mapping in @xmath0dft can be achieved through the following criteria : compare the ground states of the true and ks system energetically such that it can account for the most close resemblance of the densities in a least square sense .
sh showed it by comparing the expectation value of the original ground - state ks hamiltonian ( obtained using the harbola - sahni @xcite exact exchange potential ) with that of the alternative ks systems .
finally , the kinetic energy of true and ks system need to be kept closest .
this is also another way of comparing the ground states based on the differential virial theorem(dvt ) @xcite ; ( iv ) the cs approach is capable of generating all the potentials for a given excited state density and at the same time uniquely establishes the density - to - potential mapping .
the work is organized as follows . in sec.ii
, the generalized / unified cs @xmath0dft will be briefly discussed from the prospective of density - to - potential mapping .
it will be shown that there exist multitude of potentials for a given density .
in sec.iii , furtherance of sh @xmath0dft will be presented in order to show the existence and unique construction of the desired density - to - potential mapping . in this
, we will show , how the proposed @xmath0dft is also consistent with the generalized adiabatic connection(gac ) ks formalism @xcite and in principle applicable to ( non-)coulombic densities . in sec.iv
, we will show the existence of multiple potentials for given ground or lowest excited states can never be ruled out even within li .
et al.@xcite demonstration of gunnarsson and lundqvist(gl ) theorem @xcite . however , based on the theories presented in sec.ii & iii ,
these seemingly contradictory results will be explained in order to justify the non - violation of hohenberg - kohn(hk ) @xcite and gl theorems .
thus the density - to - potential mapping will be demonstrated within @xcite approach by making use of the unified @xmath0dft for the two model systems ( i.e. @xmath3 quantum harmonic oscillator(qho ) with finite boundary and infinite well external potentials ) . for completeness , in sec .
v , same set of model systems will be used to exemplify density - to - potential mapping based on the cs formalism @xcite .
finally , we will provide firm footing to density - to - potential mapping based on the proposed criteria of @xmath0dft .
although in principle the ground - state cs formalism @xcite has all the information about the excited - states , the desired density - to - potential mapping for individual excited - states are not so trivial and straightforward .
to do so , series of attempts being made based on the original cs approach @xcite . in the recent past , the form of functional for ground state ( both for degenerate and non - degenerate ) has been extended @xcite to study the excited states .
now we will briefly describe how the generalized cs formalism explains the existence of multiple potentials for any given fermionic density without hindering the density - to - potential mapping .
let s consider @xmath4 fermions trapped in a local external potential @xmath5 , described by the hamiltonian @xmath6 = \hat t + \hat v_{\text{ee } } + \sum_{i=1}^n\hat v_{\text{ext}}(\vec r_i ) , \label{gcs1}\ ] ] where @xmath7 and @xmath8 are the kinetic and electron - electron interaction operators with the corresponding stationary states are given by @xmath9 \psi_k(\vec r ) = e_k[\hat v(\vec r ) , n ] \psi_k(\vec r)~ , \label{gcs2}\ ] ] where @xmath10 . in eq.([gcs2 ] ) , @xmath11 $ ] are the pure state @xmath12representable stationary quantum states i.e. it is coming from the solution of the schrdinger equation . but for @xmath13representable densities ( i.e. @xmath14 ) and therefore wavefunctions
( i.e. @xmath15}^2 d\vec r = n$ ] ) , similar to the hk universal functional there exists an analogous functional which is stationary w.r.t all the variations that do not change the density ( i.e. @xmath16 ) and is given by @xmath17 = \delta_{\psi[n ] \to \rho(\vec r)}\langle\psi | \hat t + \hat v_{\text{ee } } | \psi\rangle~. \label{gcs3}\ ] ] now according to the perdew and levy extremum principle @xcite and generalized cs formalism @xcite , the energy of the @xmath18 excited state is given by @xmath19 = q^s[\rho_k;n ] + \int\rho_k(\vec r ) { v}_{\text{ext}}(\vec r ) d\vec r .
\label{gcs4}\ ] ] in eq.([gcs4 ] ) , the minimization occurs only over grling s stationary - state functional @xmath20 $ ] and the corresponding wavefunctions are given by @xmath21 = \arg\min_{\psi[n ] \to \rho_k}\langle\psi[n]|\hat t + \hat v_{\text{ee}}| \psi[n]\rangle . \label{swfc}\ ] ] on the other hand , in the ln @xcite variational constrained minimization approach for excited - states leads to the @xmath18 stationary state energy @xmath22 & = & \min_{\rho[\hat v ] \to
n}\big\{\int \rho(\vec r ) { v}_{\text{ext}}(\vec r ) d\vec r + f[\rho,\rho_0]\big\ } \nonumber\\ & = & \int \rho_k(\vec r ) { v}_{\text{ext}}(\vec r ) d\vec r + f[\rho_k,\rho_0]~ , \label{gcs5}\end{aligned}\ ] ] where @xmath23 is the ground state density of the system under consideration .
the ln energy density functional differs from the hks ground - state and the stationary state @xmath0dft functional due to the bifunctional @xmath24 $ ] , which is defined by @xmath25 & = & \min_{\psi[n ] \to \rho , { \langle\psi[n]|\psi_j[\hat v;n ] \rangle=0 , j < k } } \langle\psi|\hat t + \hat v_{\text{ee}}|\psi\rangle \nonumber \\ & = & f[\rho_k,\rho_0 ] , \label{gcs6}\end{aligned}\ ] ] for the @xmath18 excited state .
so the energy of the @xmath18 excited state can be re - expressed as @xmath22 = \min_{\rho[\hat v ] \to n } \big\{\int \rho_k(\vec r ) { v}_{\text{ext}}(\vec r ) d\vec
r + \nonumber\\ \min_{\psi[n ] \to \rho , { \langle\psi[n]|\psi_j[\hat v;n ] \rangle=0 , j < k } } \langle\psi| \hat t + \hat v_{\text{ee}}|\psi\rangle\big\}~ \label{gcs7}\end{aligned}\ ] ] with the minimizing wavefunction denoted by @xmath26 = \arg \min_{\psi[n ] \to \rho , { \langle\psi[n]|\psi_j[\hat v;n ] \rangle=0 , j < k } } \langle\psi|\hat t + \hat v_{\text{ee}}|\psi\rangle . \label{vwfc}\ ] ] in the ln bifunctional @xmath24 $ ] , if @xmath27 then the functional reduces to hk universal functional and the same holds true for @xmath28 $ ] . also @xmath24 $ ] is the generalization of the @xmath0dft stationary state functional @xmath28 $ ] as described in the @xmath29 of @xcite .
these theorems are in fact an artifact of the orthogonality constraint .
since all the lower states @xmath30 ( j < k)$ ] are determined from the external potential @xmath31 ( which is a unique functional of ground state density @xmath23 according to hk @xcite theorem ) , implies that the ground state density plays an important role in ln - formalism .
so , in principle one can also write the excited - state density bifunctional as @xmath32 $ ] instead of @xmath33 $ ]
. if the electronic densities are @xmath12representable then eq.([gcs6 ] ) modifies to @xmath34 = \int \rho_k(\vec r ) { v}_{\text{ext}}(\vec r ) d\vec r + f_k[\rho,{\hat v}_{\text{ext } } ] .
\label{gcs8}\ ] ] from the generalized cs energy functionals for any eigendensity given by the eq.([gcs4 ] ) & eq.([gcs8 ] ) there exist multiple potential functions @xcite . in general ,
these generalized multiple local external potentials can be obtained through the euler lagrange equation @xmath35 = 0 \\ v_{\text{ext}}(\vec r ) = \mu - \big(\frac{\delta f[\rho,\rho_0]}{\delta\rho}\big){\big |}_{\rho = \rho_k } \\ { \text or}~~ w_{\text{ext}}(\vec r ) = \mu - \big(\frac{\delta q^s[\rho(\vec r)]}{\delta\rho}\big){\big |}_{\rho = \rho_k } , \label{gcs9}\end{aligned}\ ] ] where @xmath36}{\delta\rho } \right)_n$ ] .
the actual potential is one of these which should be uniquely mapped to the given density as will be shown in the following sections . in particular
, the local external potentials will be identical @xmath37 iff the density @xmath38 is pure state @xmath39-representable and @xmath40 is the corresponding eigen energy .
now to obtain the ks like equation for the generation of @xmath41 and to obtain @xmath40 , one needs to first construct a non -interacting system with some external potential @xmath42 such that it s @xmath43 excited state density @xmath44 ( say ) may be the same as @xmath38 of the original system @xmath45 . in stationary - state @xmath0dft @xcite , this is done by generalized adiabatic connection ( gac ) @xcite .
whereas , in ln variational @xmath0dft @xcite , this is done by the constrained minimization of the expectation value @xmath46|\hat t + { \{{\hat v}_{\text{ee}}=0\}}|\psi[{\hat v}'_{\text{ext}},\rho^{{\hat v}'_{\text{ext}}}_m(\vec r)]\rangle$ ] , where @xmath47 $ ] gives the desire density of interest . out of many such non - interacting @xmath48 $ ]
s ( different systems ) , the unique one is chosen whose ground - state density @xmath49(say ) resembles with the ground - state density @xmath50 of the original system most closely in a least - square sense"(i.e .
the ln criterion ) .
the matching of the ground - state densities actually matches the external potentials @xmath42 and @xmath51 according to the hk theorem @xcite .
but the difference occurs between the kinetic energies of the two systems . as matter of which
, the discrepancy in the @xmath52 mapping arises because the ln criterion strictly depends upon the behavior of the bifunctional .
the cs formulation described in the previous section implies that the content of the excited state functionals @xmath53 $ ] and @xmath54 $ ] differs from the hk universal functional @xmath55 $ ] except their stationarity with respect to variation in the external potential .
actually , only in the case of ground - state , all the three functionals are identical to one another and in general there exists a close link between grling @xmath53 $ ] and levy - nagy @xmath54 $ ] @xcite .
so in the unified @xmath0dft formalism , for a given excited - state eigendensity @xmath56 , both @xmath53 $ ] and @xmath57 $ ] are stationary about the corresponding @xmath58 which also holds for the desired excited - state @xmath59 @xcite .
now due to the presence of orthogonality constraint in @xmath57 $ ] , several choices for the set of low lying states can be made to which @xmath60 will be orthogonal and for each choice , there may exists a generalized potential function @xmath61 .
so some extra deciding factors are required for setting up the @xmath52 mapping which is the intent of the current section .
now resorting back to the work of samal - harbola @xcite , we would also like to re - emphasis that the direct or indirect comparison of ground states are not sufficient to establish the @xmath62 mapping or to construct the ks system for excited - states @xcite . given the discussions on unified cs @xmath0dft in the previous section , we now present a consistent approach to address the density - to - potential mapping issues .
fundamentally rigorous and crucial tenets of the proposed dft are : ( _ i _ ) there exist ways for mapping an excited - state density @xmath56 to the corresponding many - electron wavefunction @xmath63 which in turn maps to the external potential @xmath64 through the @xmath65-stationary wavefunctions @xcite . in this
, the wavefunction depends upon the ground - state density @xmath23 implicitly .
( _ ii _ ) the ks system is to be defined through a comparison of the kinetic energy , ground - state density and variation of the energy w.r.t .
symmetry of the excited - states .
the claim is , unified cs approach can provide the mapping from an excited - state density @xmath56 to many - body wavefunction .
stationary state formalism @xcite provides a straightforward method of mapping @xmath66 , just by making sure whether @xmath67 is stationary or not , subject to the condition that @xmath68 gives @xmath69 .
but @xcite shows that different @xmath70s correspond to potentials @xmath71 .
the same problem also pervades through the variational @xmath0dft approach as proposed by ln @xcite .
thus unified cs gives , many different wavefunctions @xmath70 and the corresponding external potential @xmath71 can be associated with a given density .
now if in addition to the excited - state density we also have the ground - state information @xmath23 , then @xmath72 can be uniquely determined out of all possible multiple potentials @xmath71 . hence with the knowledge of @xmath23 , it is quite trivial to select a particular @xmath73 that belongs to a given @xmath74 $ ] combination by comparing @xmath71 with the actual @xmath72 .
alternatively , one can think of it as finding @xmath73 variationally for a @xmath75 $ ] combination .
its because the knowledge of @xmath23 and @xmath76 is equivalent . now with the above information , the bifunctional @xmath77 $ ] can be redefined as @xmath78 \ ; = \ ; \langle \psi[\rho_e,\rho_{0 } ] | \hat{t } \;+\ ; \hat{v}_{ee } | \psi[\rho_e,\rho_{0}]\rangle .
\label{cdpm1}\ ] ] the above theoretical formulation is similar to that of ln @xcite but avoids the orthogonality constraint imposed by ln formalism .
this is because , the densities for different excited states for a given ground - state density @xmath23 ( that corresponds to a unique external potential @xmath76 ) can be found in following manner : take a density and search for @xmath73 that makes @xmath79 stationary and simultaneously make sure whether the corresponding potential @xmath80 @xmath81}{\delta\rho}{\big|}_{\rho = \rho_e } \big)$ ] resembles the given @xmath23 ( or @xmath76 ) ; if not , search for another density and repeat the procedure until the correct @xmath65 is found .
thus it is clear that excited state orbitals @xmath73 are now functional of @xmath82 $ ] .
so the correct density @xmath65 is excited state density of the potential and the @xmath73 obtained in this method is also excited state wavefunction corresponding to that potential and density . after finding the correct density @xmath69 ,
make a variation over it so that @xmath83 and again perform the cs to find @xmath84 $ ] . in this case
, choose that @xmath85 which converges to @xmath45 as @xmath86 .
the above propositions for the excited - states in terms of their densities are quite reasonable , particularly because it s development is parallel to that for the ground - state dft . on the other hand ,
to construct a kohn - sham @xcite system for a given density is not so trivial ; and to carry out accurate calculations for excited - states , it is of prime importance to construct a ks system .
further , a ks system will be meaningful if the orbitals involve in an excitation match with the corresponding excitations in the true system .
samal - harbola @xcite have shown that the ks system constructed using the levy - nagy criterion fails in this regard . but using the form of the functional above a ks system can be defined for excited state .
actually , the state dependence of the excited - state exchange - correlation functional @xcite leads to the discrepancies while one compares the ground - states either direct or indirect manner .
but in principle , obtaining a ks system is plausible .
now by defining the non - interacting kinetic energy @xmath87 $ ] and using it to further define the exchange - correlation functional as @xmath88 = f[\rho_e,\rho_0 ] - e_{\text{h}}[\rho_e ] - t_s[\rho_e,\rho_0 ] , \label{cdpm4}\ ] ] solves the purpose .
so the euler equation for the excited - state densities becomes @xmath89}{\delta\rho(\vec r ) } + { v}_{\text { h}}[\rho_e ] + \frac{\delta e_{xc}\left[\rho_e,\rho_0\right]}{\delta\rho(\vec r)}\big\}~ , \label{cdpm5}\ ] ] which is equivalent to solving the ks equation @xmath90 where @xmath91 - t_{s}\left[\rho,\rho_0 \right]\big\}}{\delta\rho(\vec r)}{\big|}_{\rho(\vec r ) = \rho_e[v_{ext}(\vec r ) ] } ~. \label{cdpm7}\ ] ] in ground state dft , one can easily find the @xmath92 $ ] by minimizing the kinetic energy for a given density ; here @xmath92 $ ] for a given density is obtained by occupying the lowest energy orbitals for a non - interacting system .
but in @xmath0dft , to define @xmath87 $ ] is not easy , as for the excited - states it is not clear which orbitals to occupy for a given density . particularly because a density can be generated by many different configurations of the non - interacting systems .
levy - nagy select one of these systems by comparing the ground - state density corresponding to the excited - state non - interacting system with the true ground - state density
. however , ln criterion is not satisfactory as pointed out by samal and harbola @xcite .
the reason of this discrepancy is due to the inconsistency of the ground - state density of an excited state ks system with the true ground - state density .
the ground -state density corresponding to the excited - state ks system is not same as the ground - state density of the true system .
this means the desired state is not associated with @xmath93 , rather it comes from a different local potential @xmath94 . to settle this inconsistency ,
ks system must be so chosen that it is energetically very close to the original system and it can be ensured through the following criterion .
_ criterion i : the non - interacting kinetic energy _ @xmath95 $ ]
_ obtained through the cs need to be very close to the actual _ @xmath96 $ ] , where @xmath95 $ ] and @xmath96 $ ] are defined as @xmath97&=&\min_{\phi \to \rho_e } \langle\phi|\hat t
+ \underbrace{\hat v_{\text{ee}}=0}|
\phi\rangle\nonumber\\ t[\rho_e,\rho_0]&=&\min_{\psi \to \rho_e}\langle\psi|\hat t + \hat v_{\text{ee } } |\psi\rangle .
\label{cdpm8 } \end{aligned}\ ] ] so defining @xmath98 smallest not only ensures that dft exchange - correlation energy remains closer to the conventional quantum mechanical exchange - correlation energy but also keeps the structure of the ks potential appropriate for the desired excited - state which is shown below . based on the dvt @xcite
, it can be argued how for a given density @xmath69 one can have different exchange -correlation @xmath99 and external @xmath45 potentials . according to dvt , the exact expression for the gradient of the external potential ( for interacting system ) for a given excited - state density @xmath69 is @xmath100\vargamma_2(\vec r,\vec r')d\vec r'~ , \label{dvt1 } \label{cdpm9}\end{aligned}\ ] ] where @xmath101 .
this equation represents an exact relation between the gradient of the external potential @xmath45 , the @xmath102 interaction potential @xmath103 and the density matrices @xmath1 , @xmath104 and @xmath105 .
the vector field @xmath106 in eq.([cdpm9])is related to the kinetic - energy density tensor via @xmath107 = \big[\frac{1}{4}\big(\frac{\partial^2 } { \partial r'_\alpha\partial r''_\beta } + \frac{\partial^2}{\partial r'_\beta\partial r''_\alpha } \big)\vargamma_1(\vec r';\vec r'')\big]_{\vec r'=\vec r''=\vec r } \label{cdpm10}\ ] ] so , @xmath106 can be called a local functional of @xmath108 . similarly , for ks potential eq.([cdpm9 ] ) reduces to @xmath109 as a given ground - state density @xmath23 fixes the external potential uniquely via hk theorem , which implies that @xmath65 , @xmath108 and @xmath110 are also fixed from eq.([cdpm9 ] ) . since the density matrices generated by some eigenfunction @xmath73 of the hamiltonian @xmath111
so the fixed pair of excited - state and ground - state density i.e. @xmath82 $ ] may be arising from different configurations @xmath112 different configurations can be thought of as arising from different external potential or different exchange - correlation potential and this is due to the different @xmath108 and @xmath110 for a fixed @xmath69 .
suppose a given density @xmath69 is generated through an @xmath113 ks system , then @xmath114 if the density is generated through a @xmath115 external potential then @xmath116\vargamma^j_2(\vec
r,\vec r')d\vec r'~. \label{cdpm13}\end{aligned}\ ] ] as a matter of which @xmath117[\rho_e(\vec r ) \rho_e(\vec r ' ) - \vargamma_2(\vec r , \vec r ' ) ] d\vec r'}{\rho_e(\vec r ) } \label{cdpm-1}\end{aligned}\ ] ] becomes @xmath118 where @xmath119 is the field due to the fermi - coulomb hole of the @xmath115 system @xmath120 $ ] .
so the kinetic energy difference between the true system and ks system is given by @xmath121 \big ) -
\vec z\big(\vec r;[\vargamma_1]\big)\big\ } d\vec r . \label{cdpm15}\ ] ] this difference should be kept the smallest for the true ks system so that it gives the ks system consistent with the original system . as a matter of which , we conclude that one way to establish the @xmath122 mapping via the ln formalism @xcite is : if among the several potentials @xmath112 which have the same excited -state density , one can choose the correct ks potential by comparing the ground - state density i.e. keep that ks - potential whose ground - state density resembles with the true ground - state density . keeping the ground - state density close we actually keep the external potential fixed via hk theorem .
thus ln criterion is exact for non - interacting system as there is no interaction , so the ground - state density match perfectly .
this proposal of ln for @xmath122 mapping was carried by samal and harbola @xcite but they argued in a slightly different way .
they proposed that both for interacting and non - interacting case among all the multiple potentials , choose the correct ks potential whose ground - state density differ from the exact ground - state `` most closely by least - square sense '' which is done in the following manner .
if @xmath123 is the exact ground state density and @xmath124 is that of the ks system ( or the alternate potentials @xmath125 ) then sh proposition can be further improved intuitively .
_ criterion ii : the mean square distance between @xmath123 and @xmath124 should remain very close to zero_. thus @xmath126 = \min_{v[\tilde{\rho_0},\rho_e ] } \left\{\int _
\infty\left|[\rho_0(\vec r ) - \tilde{\rho_0}(\vec r)]\right|^2 d \vec r \right\}^\frac{1}{2 } \geq 0~ , \label{cdpm16}\ ] ] where the integration is carried out in the sobolev space .
this criterion is more appropriate in the context of @xmath122 than the one proposed by @xcite .
the criterion as given in eq.([cdpm16 ] ) will be fully satisfied if one makes use of the excited state functionals @xcite .
otherwise it may fail in certain situations as pointed out by sh @xcite . instead of sticking to the _ criterion i & ii _
, one can even go beyond the same through _ criterion iii : compare the ground states of the true and alternate systems energetically_. it can be done in the following manner in order to select the ks system for a given density . the alternative approach is to compare the ground - state expectation value of the ks system and the true system , instead of comparing their ground - state densities and kinetic energies .
the procedure for comparing ground - state energy level is the following .
first solve the exact dft equation ( say harbola - sahni @xcite etc ) for ground - state of the true system and obtain the ground - state of ks hamiltonian @xmath127 . if the expectation value of the ground state hamiltonian of the true system is @xmath128 and that of the ks system is @xmath129 , then one need to choose that ks system whose @xmath130 @xmath131 @xmath128 .
these criteria are well connected to the gac - ks @xcite as discussed below .
since gac - ks in principle helps for the self - consistent treatment of excited states and could be considered as a plausible extension of hk theorem to the same .
so now the furtherance of the propositions made by sh @xcite as discussed previously will be justified within the gac - ks .
indeed , relying on the principles of gac - ks , unified cs formalism along with the sh criteria can also establish the density - to - potential mapping at the strictly correlated regime which will be shown below . in gac , the @xmath132
dependent hamiltonian which is also used in the pl extremum principle is given by @xmath133 = \hat t + \lambda \hat v_{ee } + \sum_{i=1}^n \hat v(\vec r_i ) , \label{gaceq1}\ ] ] with the corresponding equation of state @xmath134 \psi_\lambda[\hat v , n ] = e_\lambda[\hat v , n ] \psi_\lambda[\hat v , n ] , \label{gaceq2}\ ] ] where @xmath132 is the coupling constant with @xmath135 allowing the electron - electron interaction to be triggered .
unlike the adiabatic connection ( ac)-dft , the external potential @xmath2 , is independent of @xmath132 .
analogous to the levy - lieb cs functionals , the gac for the conjugate density functionals @xmath136 $ ] ( density fixed ac ) and @xmath137 $ ] ( potential fixed ac ) are given by @xmath138 = f_{\lambda = 0}[\rho ] + \int_0 ^ 1 \frac{df_\lambda[\rho]}{d\lambda}~d\lambda ~ , \label{gaceq3}\ ] ] @xmath139 = e_{\lambda = 0}[\hat v ] + \int_0 ^ 1 \frac{de_\lambda[\hat v]}{d\lambda}~d\lambda ~. \label{gaceq4}\ ] ] similar to eq.([gaceq3 ] ) and ( [ gaceq4 ] ) , the excited - state functionals @xmath140 $ ] , @xmath141 $ ] , @xmath142 $ ] and @xmath143 $ ] can be defined . upon finding these @xmath0dft functionals , one can define the gac by starting at a @xmath65 stationary wavefunction for @xmath144 and then by gradually turning off ( @xmath145 ) the electron - electron interaction .
thus the @xmath65-stationary wavefunctions for @xmath135 will form the gac in @xmath0dft .
since the @xmath65-stationary wave functions for a given @xmath65 are numerable and the adiabatic connections do not overlap with each other , states @xmath146 of non - interacting model systems equals to the @xmath65-stationary wave functions at @xmath145 ( i.e.@xmath147 $ ] ) and can be assigned to real electronic states @xmath148 $ ] @xcite .
these assigned model states are the eigenstates of the gac - ks formalism .
as discussed above , they are eigenstates of a hamiltonian operator with local multiplicative potential . in this way , the gac will define the path of going from a non- interacting system to an interacting system via a @xmath149stationary path .
although for each of the interacting system , one can still end up with multiple non - interacting ks system .
but with the criteria discussed previously it s possible to select the appropriate ones .
so once the @xmath150 for the interacting system is fixed , it do carries over to the ks system via gac and vice versa .
this shows how the proposed unified cs formalism not only establishes the density - to - potential mapping concretely but also constructs the ks system successfully . in the following sections we will exemplify what we have proposed so far through two model systems
this will be done in order address the critiques about density - to - potential mapping in @xmath0dft .
the issue of non - uniqueness in the density - to - potential mapping is also persuaded @xcite in the context of gl theorem @xcite . in @xcite , it has been demonstrated for higher excited states of the considered @xmath3 model system there is no equivalence of the gl / hk theorem .
but the critical analysis of @xcite presented in this section will outline how the multiplicity of potentials still ca nt be ruled out even in the case of ground as well as lowest excited states .
so one need to go beyond @xcite approach in order to address the validity of hk & gl theorem for such state .
in fact , relying on the principles of unified @xmath0dft approach @xcite as discussed in sec .
ii along with the proposed criteria of sec .
iii , it will be shown here why the claim made in @xcite lacks merit to address the excited - state density - to - potential mapping . to validate the density - to - potential mapping ( i.e. the analogue of hk / gl theorem ) in @xcite proposed approach
, we will consider as test cases : the examples of the @xmath3 qho with finite boundary and then the infinite potential well .
for clarity in understanding let s first briefly discuss the theoretical formulation of @xcite .
the schrdinger equation of two non - interacting fermions subjected to local one dimensional potentials @xmath151 and @xmath152 s.t .
@xmath153 , where @xmath154 is a constant are given by @xmath155\phi_i(x ) = \varepsilon_i \phi_i(x)~ , \label{gl1}\ ] ] @xmath156\psi_i(x ) = \lambda_i \psi_i(x)~. \label{gl2}\ ] ] suppose that the eigenfunctions of the local potential @xmath152 generates the ground / excited - state eigendensity of @xmath151 as one of it s eigendensity but with some arbitrary configuration which is either same or different from the original one .
then one possible way of achieving this is : the wavefunctions @xmath157 of the potential @xmath152 can be associated to the wavefunctions @xmath158 of the potential @xmath151 via the following unitary transformation i.e. @xmath159 as a matter of which the density preserving constraint will be satisfied and the ground / excited state density of two potentials remain invariant i.e. @xmath160 now the potentials can be obtained from the eqs.([gl1 ] ) and ( [ gl2 ] ) by inverting the same @xmath161 also from eqs.([gl1 ] ) and ( [ gl2 ] ) , the difference between any two eigenvalues @xmath162 and @xmath163 corresponding to the potentials @xmath151 and @xmath152 are given by @xmath164~ , \label{gl6}\ ] ] @xmath165~. \label{gl7}\ ] ] now by plugging the values @xmath166 and @xmath167 from eq.([gl3 ] ) back in eq.([gl7 ] ) , the rotation @xmath168 can be obtained from the following @xmath169\\\\ = \delta'[2\phi_i(x)\phi_j(x\cos2\theta(x)+\{\phi^2_j(x)-\phi^2_i(x)\}\sin2\theta(x ) ] \label{gl8 } \end{array}\ ] ] or @xmath170 where @xmath171~. \label{gl10}\end{aligned}\ ] ] the eq.([gl9 ] ) is the central equation of @xcite theoretical framework which need to be solved numerically with proper initial conditions in order to obtain the alternate potential @xmath152 for any given density and eigenvalue differences . in this work ,
the adopted numerical procedure to solve the above mentioned differential equation is very much accurate even at the boundary where obtaining appropriate structure and behavior of the multiple potentials and the corresponding wavefunctions are important and crucial . the first model system for demonstrating the density - to - potential mapping is the @xmath3 qho defined by @xmath172 so the wavefunctions and energy eigenvalues of the @xmath173 eigenstate are given by @xmath174 where @xmath175 + ( atomic units are adopted i.e. @xmath176 and @xmath177 ) ( blue ) and @xmath178 ( green ) for @xmath179 .
lower panel : shows the alternate potential associated with above wavefunctions and density.,width=240,height=288 ] but with @xmath180.,width=240,height=288 ] now consider two non - interacting fermions occupying the ground - state of the qho i.e. @xmath181 .
so the eigenvalue difference for this state @xmath182 and the density is given by @xmath183 thus the corresponding equation for rotation @xmath168 can be obtained from the eq.([gl9 ] ) and is given by @xmath184 = 0~. \label{qhoeq4}\ ] ] now eq.([qhoeq4 ] ) has to be solved with proper initial conditions .
the initial conditions can be fixed by taking into consideration the symmetry of the differential eq.([qhoeq4 ] ) and the normalization condition of the wavefunction . from eq.([qhoeq4 ] ) it is clear that @xmath185 as both @xmath158 and @xmath186 are symmetric about @xmath187 .
now another condition is that @xmath166 and @xmath167 must also be normalized .
so if we plot the renormalization @xmath188 @xmath189 as a function of @xmath190 , then the points where @xmath191 corresponds to the normalization of @xmath166 and @xmath167 @xcite and it will provide the initial condition on @xmath190 . after finding @xmath168 , the transformed set of normalized wavefunctions @xmath166 and @xmath167 is being obtained . again using these wavefunctions
the potential @xmath152 can be determined from the eq.([gl5 ] ) . in fig.[qhofig0_0 - 1 ] and fig.[qhofig0_0 - 2 ] , we have shown two different potentials which are obtained for the eigenvalue differences @xmath179 and @xmath180 respectively along with the corresponding wavefunctions .
the important point of observation here is that the ground - state density @xmath192 $ ] now corresponds to some arbitrary excite - state having density @xmath193 $ ] . as a matter
of which , for the fixed @xmath194 and @xmath195 , the system gets transformed to some other system @xmath152 for which @xmath196=\rho_0^{qho}]$ ] and/or @xmath197=\rho_0^{qho},\tilde{\rho_0}]$ ] will be stationary .
the corresponding stationary states are basically the transformed wavefunctions which are given by eq.([gl2 ] ) @xmath198 = \psi_l[\rho_e,\tilde{\rho_0}]$ ] . in this @xmath199
is the ground state density of the newly generated potential @xmath152 and @xmath200 .
now from the proposed criteria it follows that @xmath201 , @xmath202 > 0 $ ] and new system is energetically far off from the original one .
hence the given ground - state density should be uniquely mapped to the original qho potential @xmath151 although there exist several multiple potentials @xmath152 .
this result is consistent with the generalized / unified cs formalism @xcite .
but for the lowest excited state density being produced with @xmath203.,width=240,height=288 ] but with @xmath204.,width=240,height=288 ] as the second example , we consider the lowest excited - state of the qho .
so the two non - interacting fermions are now occupying the @xmath205 and @xmath206 state .
for this case , @xmath207 , @xmath208 and @xmath209 with the density @xmath210 and the corresponding equation for rotation @xmath168 is given by @xmath211 = 0~. \label{eq34}\end{aligned}\ ] ] since in this case @xmath212 is symmetric , @xmath213 is antisymmetric , so @xmath186 symmetric around @xmath187 .
thus eq.([eq34 ] ) implies that @xmath168 should be symmetric at @xmath187 .
the initial conditions on @xmath214 is obtained from the behavior of the renormalization @xmath188 as a function of @xmath214 .
following the same procedure as before , in this case also we have obtained different potentials for the fixed lowest excited state density which are shown in the fig.[qhofig0_1 - 1 ] and fig.[qhofig0_1 - 2 ] .
these two alternative potentials and the transformed wavefunctions correspond to two different eigenvalue differences @xmath215 and @xmath204 . as described in the ground state case , in this case also the structure of the potential is different from the original 1d qho as the potential should follow the structure of the wavefunctions . however , according to the unified cs @xmath0dft , the results are never due to the violation of gl theorem .
this is because the ground and lowest excited states of the newly found potential are quite different from that of the qho .
so following similar argument as in the previous case , now the lowest excited - state density of the qho corresponds to some different eigendensity of the multiple potentials .
thus the multitude of potentials poses no issues for the validity of the gl theorem .
but for one of the higher excited state density being produced with @xmath216.,width=240,height=288 ] but with @xmath217.,width=240,height=288 ] and @xmath178 ( green & red ) resulting the ground state density of 1d potential well for @xmath218 .
lower panel : shows the alternate potentials ( green & red ) and the density ( magenta ) associated with above wavefunctions.,width=288,height=336 ] but with @xmath219.,width=288,height=336 ] but with @xmath220.,width=288,height=336 ] but for the lowest excited state density being produced with @xmath218.,width=288,height=336 ] but with @xmath219.,width=288,height=336 ] but with @xmath220.,width=288,height=336 ] but for one of the higher excited state density being produced with @xmath218.,width=288,height=336 ] but with @xmath219.,width=288,height=336 ] but with @xmath220.,width=288,height=336 ] here we consider one of the higher excited - state of 1d qho ( i.e. two non - interacting fermions are in the @xmath205 and @xmath221 states ) . for this case , the eigenvalue difference is @xmath222 and the density corresponding to it is given by @xmath223 similarly , the corresponding equation for rotation @xmath168 is given by @xmath224 = 0~.\nonumber\\ \label{qho02 - 2}\end{aligned}\ ] ] now by solving eq.([qho02 - 2 ] ) for rotation @xmath168 in analogous with the ground - state of the qho and after taking care of the normalization of the transformed wavefunctions , the potential @xmath152 is obtained for @xmath225 .
the potentials along with the wavefunctions are shown in fig .
[ qhofig0_2 - 1 ] & fig.[qhofig0_2 - 2 ] .
similar to ground and lowest excited - state , here too the given density is a different eigendensity of the new potentials . if it would have the same eigendensity of @xmath152 then @xmath152 should have been identical to the @xmath226 .
but it is not the case .
that s why the generated potentials are completely different from the qho . as our second case study
, we consider the model system same as that reported in @xcite ( i.e. particles are trapped inside an @xmath3 infinite potential well ) .
for an infinite potential well with length varying from @xmath227 to @xmath228 , the @xmath173 eigenfunction @xmath229 and the energy eigenvalue @xmath230 are given by @xmath231 where @xmath232 .
the density @xmath186 corresponding to the two potentials @xmath151 and @xmath152 is given by eq.([gl4 ] ) .
for two spinless non - interacting particles in @xmath233 states , the energies of two states and the difference are @xmath234 the density corresponding to these states is @xmath235~ , \label{pbeq5}\ ] ] and the equation corresponding to eq.([gl9 ] ) for the rotation @xmath168 is @xmath236 = 0~. \label{pbeq6}\ ] ] since @xmath213 is symmetric and @xmath186 is symmetric about @xmath237 . thus eq.([pbeq6 ] ) indicates that @xmath168 should be symmetric such that @xmath238 . with this initial condition and choosing any value of @xmath163
one can solve for @xmath168 and subsequently obtain the @xmath68 s. now using these @xmath68 s , the alternate potentials @xmath152 will be obtained by using eq .
( [ gl5 ] ) . since the transformed wavefunction @xmath166 must also be normalized .
this condition will be fulfilled by choosing the appropriate value of @xmath239 at which the @xmath166 should be normalized .
once @xmath166 is normalized then @xmath167 will also be normalized . again by adopting the same procedure as that described in the case of @xmath3 qho , the alternative multiple potentials are obtained by making use of the following renormalization @xmath188 condition @xmath240 all the wavefunctions , densities and multiple potentials are shown in the figs.([pbfig2 - 1 ] @xmath241 [ pbfig2 - 3 ] ) .
here we have generated the multiple potentials for @xmath242 respectively .
as is expected , the wavefunctions are totally different from the ground state of the @xmath3 infinite well .
although the density remains to be the same in all the cases .
but its not the ground state eigendensity of the multiple potentials .
so this poses no issue for the hk theorem .
now consider two fermions occupying the @xmath243 ( i.e. the lowest excited - state ) eigenstates of the infinite potential well . here
too we have obtained several multiple potentials unlike @xcite . for this excited - state , the energy eigenvalues are @xmath244 , @xmath245 with @xmath246 . hence the density arising from these two states
is given by @xmath247~. \label{pbeq8}\ ] ] similar to the previous examples , the equation for the rotation @xmath168 is the following @xmath248 = 0~. \label{pbeq9}\end{aligned}\ ] ] here @xmath213 is symmetric , @xmath249 is antisymmetric and @xmath186 symmetric about @xmath250
. thus eq.([pbeq9 ] ) predicts that @xmath168 is antisymmetric such that @xmath251 . in this case also normalization of both @xmath166 and @xmath167 are taken care and the proper @xmath188 ( renormalization ) values are obtained w.r.t .
@xmath252 . quite interestingly , in this case also we have successfully generated multiple potentials for @xmath242 .
this is where @xcite failed to explain the validity of gl theorem .
as expected , the potential follows the wavefunctions pattern .
this is obvious at the boundary where the wavefunctions are perfectly vanishing , the potential shoots up to a very large positive value .
the potentials along with wavefunctions are shown in the figs.([pbfig3 - 1 ] @xmath241 [ pbfig3 - 3 ] ) . following the same argument as in the case of the previous model system ,
the multiplicity of potentials obtained here are nothing to do with the gl theorem . now to complete our exploration on @xmath3 well , we have considered here the second excited - state of it .
this is the only excited - state for which @xcite reported multiple external potentials for various eigenvalue differences .
we too generated multiple potentials and the corresponding wavefunctions for @xmath242 which are shown in the figs.([pbfig4 - 1 ] @xmath241 [ pbfig4 - 3 ] ) .
the results follow the trend similar to that of the ground and lowest excited - state . in all the cases ,
we have noticed that the potentials and the corresponding rotation angles can never attain flat structure at the boundary unlike @xcite .
$ ] is the excited - state density of @xmath3 potential well with ground - state @xmath253 .
@xmath254 is the ground - state density of potential @xmath255 whose excited - state configuration @xmath256 $ ] results the same @xmath69 .
( b ) @xmath257 $ ] is the potential whose ground - state configuration results the same @xmath69 of ( a ) and is shown along with @xmath258 $ ] .
( c ) @xmath259 $ ] is the excited - state density of @xmath3 potential well with ground - state @xmath23 and produced in an alternative configuration @xmath260~(v_1[\tilde{\rho}^{(1)}_0])$ ] besides the ground - state configuration leading to @xmath257 $ ] .
( d ) shows all the alternative potentials of ( c).,width=316,height=288 ] $ ] is the excited - state density of @xmath3 potential well with ground - state @xmath253 . @xmath254 and @xmath261 are the ground -state densities of @xmath255 and @xmath262 whose excited - state configurations @xmath256 $ ] and @xmath260 $ ] results the same @xmath69 .
( b ) @xmath263 $ ] is the potential whose ground - state configuration gives the same @xmath69 of ( a ) and is shown along with @xmath255 , @xmath262 .
( c ) @xmath264 $ ] is the excited - state density produced in alternative configuration @xmath256~(v_1[\tilde{\rho}^{(1)}_0])$ ] , besides the ground - state configuration leading to @xmath257 $ ] .
( d ) shows all the alternative potentials in ( c).,width=316,height=288 ] $ ] is the excited - state density of @xmath3 infinite potential well with ground - state @xmath253 . @xmath254 and @xmath261 are the ground - state densities of @xmath255 and @xmath262 , whose excited - state configurations @xmath256 $ ] and @xmath265 $ ] results the same @xmath69 .
( b ) @xmath266 is the potential whose ground - state density is same as @xmath69 of ( a ) and is shown along with @xmath255 , @xmath262 .
( c ) @xmath267 $ ] is the excited - state density produced via the alternative configurations @xmath268~(v_1[\tilde{\rho}^{(1)}_0])$ ] besides the ground - state configuration leading to @xmath257 $ ] .
( d ) shows all the alternative potentials of ( c).,width=316,height=288 ] $ ] is the excited - state density of @xmath3 infinite potential well with ground - state @xmath253 .
@xmath254 , @xmath261 and @xmath269 are the ground - state densities of @xmath255 , @xmath262 and @xmath266 whose excited - state configurations @xmath256$],@xmath260 $ ] and @xmath270 $ ] results the same @xmath69 .
( b ) @xmath271 is the potential whose ground - state density is same as @xmath69 of ( a ) and is shown along with @xmath255 , @xmath262 and @xmath266 .
( c ) @xmath272 $ ] is the excited - state density produced in the alternative configurations @xmath256~(v_1[\tilde{\rho}^{(1)}_0])$ ] , @xmath260~(v_2[\tilde{\rho}^{(2)}_0])$ ] and @xmath265~(v_3[\tilde{\rho } ^{(3)}_0])$ ] besides the ground - state configuration leading to @xmath273 $ ] .
( d ) shows all the alternative potentials of ( c).,width=316,height=288 ] $ ] is the excited - state density of @xmath3 infinite potential well with ground - state @xmath253 .
@xmath254 and @xmath261 are the ground state densities of @xmath255 and @xmath262 , whose excited - state configurations @xmath256 $ ] and @xmath260 $ ] results the same @xmath69 .
( b ) @xmath266 is the potential whose ground state density is same as @xmath69 of ( a ) and is shown along with @xmath255 , @xmath262 .
( c ) @xmath274 $ ] is the excited - state density produced in alternative configurations @xmath256 ~(v_1[\tilde{\rho}^{(1)}_0])$ ] and @xmath260~(v_2[\tilde{\rho}^{(2)}_0])$ ] besides the ground - state configuration leading to @xmath263 $ ] .
( d ) shows all the alternative potentials of ( c).,width=316,height=288 ] $ ] is the excited - state density of @xmath3 infinite potential well with ground - state @xmath253 . @xmath254 and @xmath261 are the ground - state densities of @xmath255 and @xmath262 , whose excited - state configurations @xmath275 $ ] and @xmath270 $ ] results the same @xmath69 .
( b ) @xmath266 is the potential whose ground - state density is same as @xmath69 of ( a ) and is shown along with @xmath255 , @xmath262 .
( c ) @xmath276 $ ] is the excited - state density produced in alternative configurations @xmath256~(v_1[\tilde{\rho}^{(1)}_0])$ ] and @xmath260 ~(v_2[\tilde{\rho}^{(2)}_0])$ ] besides the ground - state configuration leading to @xmath263 $ ] .
( d ) shows all the alternative potentials of ( c).,width=316,height=288 ] $ ] ( both half - filled ) , @xmath277 $ ] ( both half filled ) and @xmath278$](both half filled ) are the excited - state densities of the potential @xmath279 produced as the ground state density of the potentials @xmath255 , @xmath262 and @xmath266 .
( b ) shows all the four potentials @xmath279 , @xmath255 , @xmath262 and @xmath266 of ( a ) .
( c ) @xmath280 $ ] ( both half filled ) is the excited - state density of the potential @xmath279 produced in an alternative excited state configuration @xmath281~(v_1)$ ] .
( d ) shows both the potentials of ( c).,width=288,height=288 ] to conclude this section , we would like to shed some light on the structure of the generated potentials at the boundary as its very important to be determined accurately . since the wavefunctions die out towards the boundary .
thus the potentials obtained by the schrdinger equation inversion ( i.e. eq.[gl5 ] ) for specified eigenvalue differences will attain large positive value . actually , in our approach we have gone way beyond @xcite to generate the accurate structure of the potential which is clear from the results .
the important point to be noted is that the singularity of the potential plays the most crucial role if one directly solving the schrdinger equation .
but in getting the potential structure whether by inverting the schrdinger equation or cs method solely depends on the wavefunction behavior in a given domain .
so better access of the wavefunction s behavior will by default lead to reliable potential structure .
in this section , we will discuss the results in connection with the density - to - potential mapping based on the cs - formalism discussed earlier . according to it , there exist multiple potentials for a given ground or excited state ( eigen)density .
but for the case of excited state density , when it is produced as some different excited - state of these multiple potentials ( except the actual one ) the corresponding ground - states are completely different from that of the original system .
similarly , one can produce potentials whose ground - state density may be same as the excited - state density of the original system .
the results we have obtained for the systems of our study are fully consistent with the unified cs @xmath0dft .
the zhao - parr @xcite cs method is being used to show the multiplicity of potentials for a given density . to begin the cs exemplification ( shown in fig.[csfig1 ] ) , lets consider @xmath282 non - interacting particles in an @xmath3 potential well , where two fermions are in @xmath283 state and one fermion each in @xmath284 and @xmath285 state . as a result
, this gives some excited state density @xmath286 associated with the above configuration which is shown in the fig.[csfig1](a ) and is given by @xmath287 where @xmath288 s are the wavefunctions of the @xmath3 potential well . in all our results shown in the figures ( [ csfig1 ] ) to ( [ csfig7 ] ) , we have adopted notation @xmath289 , where @xmath290 denotes the quantum number of the eigenfunctions of the potential @xmath279 or @xmath291 ( @xmath292 ) and @xmath293 , the occupation .
using cs @xcite the excited state density @xmath286 given by eq.([cseq1 ] ) is produced through another alternative potential @xmath255 ( say ) whose @xmath283 state is occupied with @xmath294 fermions ( i.e. @xmath295 ) and @xmath296 with one fermion each ( i.e. @xmath297 ) .
now the ground state density of the potential @xmath255 is different from that of the @xmath298 ( i.e. particle in an infinite potential well ) which is given by @xmath254 ( fig.[csfig1]a ) . as per our formalism
, there can be many such multiple potentials having the given density as it s eigendensity associated with some combination of eigenfunctions .
so it is possible that one can also obtain second alternative potential @xmath262 ( say ) whose ground - state density is same as the above excited state density ( @xmath286 ) of the original system ( @xmath298 ) . in this way , we have studied six such excited states of the @xmath3 potential well ( figs.[csfig1 ] to [ csfig6 ] ) and for each case we are able to produce symmetrically different multiple potentials for fix densities .
also in each case , we have produced the alternative potential whose ground - state density is nothing but the given excited - state density of the original configuration ( i.e. @xmath3 potential well ) . as our final case study , we have considered the excited - states of the @xmath3 qho .
this is also an interesting model system like the potential well .
the results for this case , are shown in fig.[csfig7 ] .
now consider the fig.[csfig7]a , in this case we have produced three symmetrically different alternative potentials @xmath255 , @xmath262 and @xmath266 ( shown in fig.[csfig7]b ) whose ground - states densities ( i.e. @xmath299 and @xmath300)are same as the different excited - states densities ( i.e. @xmath301 and @xmath302 ) of the qho potential @xmath303 . here
@xmath304 corresponds to the configuration [ @xmath305 .
similarly , @xmath306 and @xmath302 are arising from the excited - state configurations [ @xmath307 and [ @xmath308 respectively . in fig.[csfig7](d ) , we have produced a different potential @xmath255 whose excited - state density corresponding to the configuration [ @xmath309 is same as the excited - state density @xmath286 ( [ @xmath305 ) of the original @xmath3 qho potential .
although we have produced so many potentials , but our criteria will only select the original potentials ( i.e. the infinite potential well in the previous and qho in the current study ) for any given ( i.e. either ground or excited - state ) density . thus establishes the excited - state @xmath310 mapping uniquely .
now the conceptually basic questions of @xmath0dft : what are the consequences as well as similarities and differences between the results of the cs formalism and that obtained in connection to the hk / gl theorem ? secondly , whether there arisen any critical scenario which is inconsistent with the hk and /or
gl theorem(s ) ?
this is because several multiple potentials are obtained for non - interacting fermions in the ground as well as lowest excited state . not only that , @xcite have also claimed that for higher excited - states there is no analogue of hk theorem .
so the seemingly contradictory results may give rise to the wrong conclusion about the validity of hk / gl theorem and non - existence of density - to- potential mapping for excited - states .
however , the generalized / unified cs formalism overrules all these claims by showing that the ground - state density of a given symmetry ( potential ) can be the excited - state density of differing symmetry ( potential ) .
now this excited - state will have a corresponding ground state which will be obviously quite different from the ground - state of the original system . as a matter
of which there will exist a different potential according to hk theorem .
this is also true for the excited -state density of the actual system : when it becomes either the ground - state or some arbitrary excited -state density of another potential .
so the unified cs formalism justifies the non - violation of hk / gl theorem for such states .
now based on the unified / generalized cs @xmath0dft , one can very nicely interpret ours as well as @xcite results . actually by keeping the excited / ground state density fix via a unitary transformation
never guarantee the symmetries of the states involve will remain intact .
this is because by changing the @xmath163 value and keeping either ground or the excited state density fix , we are forcing the system to change itself accordingly without hindering only the fixed density constraint . since @xmath163 is not fixed .
so in principle one can make several choices for @xmath163 and for each choice , the system will converge to different potentials ( systems / configurations ) which can give the desired density of ground/ excited - state of the original system ( potential / configurations ) as one of it s eigendensity . actually , the converged potentials are those for which the grling and ln functionals are stationary and minimum respectively .
so everything is again automatically fits into realm of generalized cs formalism and nothing really contradicting or posing issues for the @xmath0dft formulations provided by @xcite .
also the transformed quantum states leading to multitude of potential for a given density are energetically far off from the actual system and even the ground - states are also very different . thus the generalized cs formalism proposed in this work along with the sh criteria can be considered as the most essential steps for establishing the @xmath311 which further elaborated below .
now the question is out of these existing multiple potentials in association with a fix density and @xmath163 , which potential in principle should be picked in view of the @xmath312 ? the criteria of selecting the exact potential out of all possibilities
have already been discussed in sec.iii .
first of all it is quite obvious from the figs.([qhofig0_0 - 1 ] @xmath241 [ qhofig0_2 - 2 ] ) and from fig.[pbfig2 - 1 ] @xmath241 fig.[pbfig4 - 3 ] that the ground - state densities of the generated alternate potentials are different from that of the original potential .
this is also true even for the results of the cs formalism as shown in the figs.([csfig1 ] @xmath241 [ csfig7 ] ) .
so when we are fixing the excited - state density at the same time we should have taken care of the ground - state of the newly found system and the old one . similarly , when several multiple potentials are generated for a given ground - state density , the same is not produced as the ground state eigendensity of the alternate potentials .
so it s quite obvious that there is no violation of the hk theorem .
the criteria of taking care of the ground - states of the two system is given in eq.([cdpm16 ] ) .
additionally the kinetic energies of the two systems need to be kept closest , which we have pointed out on the basis of dvt .
so in all the non - interacting model systems reported here , @xmath313 should have been zero .
but the drastically differing structures of the transformed and original wavefunctions are nothing but the manifestation of non - vanishing difference of kinetic energies and thus leading to the multiple potentials .
furthermore , the most significant differences between the symmetries of the old and new systems implies that principally there exist discrepancies in the expectation values of the hamiltonian w.r.t .
the ground - states of various multiple potentials .
this is what trivially follows from the reported results .
hence , the proposed criteria uniquely maps a given density of the @xmath3 qho / infinite well to a potential which is nothing but the @xmath3 qho / infinite well and discards rest of the multiple potentials .
in this work , we have tried to obtain a consistent theory for @xmath0dft based on the stationary state , variational and gac formalism of modern dft .
we have provided a unified and general approach for dealing with excited - states which follows from previous attempts made by perdew - levy , grling , levy - nagy - ayers and in particular the work of samal - harbola in the recent past . in this current attempt
, we have answered the questions raised about the validity of hk and gl theorems to excited - states .
we have settled the issues by explaining why there exist multiple potentials not only for higher excited states but also for the ground as well as lowest excited state of given symmetry .
in fact , the existing @xmath0dft formalism allows the above possibility and at the same time keeps the uniqueness of density - to - potential mapping intact .
so we have established in a rigorous fundamental footing the non - violation of the hk and gl theorem .
actually , the generalized cs approach gives us a strong basis in choosing a potential out of several multiple potentials for a fixed ground / excited state density . in our propositions
, we have strictly defined the bi - density functionals for a fix pair of ground and excited - state densities in order to establish the density - to - potential mapping . not only that , the theory also gives us a clear definition of excited - state ks systems through the comparison of kinetic and exchange - correlation energies w.r.t . the true system .
it does takes care the stationarity and orthogonality of the quantum states .
so everything fits quite naturally into the realm of modern dft . to conclude , we have demonstrated density - to - potential mapping for non - interacting fermions .
for interacting case the gac can be used to formulate all the theoretical and numerical contents in a similar way .
we are working along this direction for strictly correlated fermions and the results will be reported in future .
finally , our conclusion is that nothing really reveals the manifestation of the failure or violation of the basic theorems and existing principles of modern dft irrespective of the states under consideration .
the method presented by samal - harbola and further progress being made here provides a most suitable framework and starting ground for the development of new density -functional methods for the self - consistent treatment of excited states .
more realistically , the unified cs @xmath0dft and further extensions to the sh criteria treat both the ground or excited states in an analogous manner .
hence , the present work endows the uniqueness of density - to - potential mapping for excited - states with a firm footing .
the authors thankfully acknowledge valuable discussions with prof . manoj k. harbola and m. hemanadhan . | the generalized constrained search formalism is used to address the issues concerning density - to - potential mapping for excited states in time - independent density - functional theory .
the multiplicity of potentials for any given density and the uniqueness in density - to - potential mapping are explained within the framework of unified constrained search formalism for excited - states due to grling , levy - nagy , samal - harbola and ayers - levy .
the extensions of samal - harbola criteria and it s link to the generalized constrained search formalism are revealed in the context of existence and unique construction of the density - to - potential mapping .
the close connections between the proposed criteria and the generalized adiabatic connection are further elaborated so as to keep the desired mapping intact at the strictly correlated regime .
exemplification of the unified constrained search formalism is done through model systems in order to demonstrate that the seemingly contradictory results reported so far are neither the true confirmation of lack of hohenberg - kohn theorem nor valid representation of violation of gunnarsson - lundqvist theorem for excited states .
hence the misleading interpretation of subtle differences between the ground and excited state density functional formalism are exemplified . |
some galaxies , despite of their simple morphologies , can reveal to be very complex sistems from a kinematic point of view .
clear examples are objects with stars and/or gas in orthogonal rotation with respect the rest of the galaxy ( e.g. , @xcite ) , or with a warped inner structure ( e.g. , @xcite ) . among these galaxies with multi - spin components ,
_ counter - rotating galaxies _ are those that host two components that rotate in opposite directions from each other .
these peculiar objects have been observed from in all morphological types .
they are classed by the nature ( stars vs. stars , stars vs. gas , gas vs. gas ) and size ( counter - rotating cores , rings , and disks ) of the counter - rotating components ( see @xcite , for a review ) .
the subject of this work is the peculiar case of galaxies hosting two counter - rotating stellar disks of comparable size .
the prototype of this class of objects is the famous e7/s0 galaxy ngc 4550 , whose counter - rotating nature was first discovered by @xcite .
the study of the kinematics and the stellar population of the two counter - rotating components are the key to understand the formation mechanisms of these objects .
this task is complicated by the fact that the two components are co - spatial , and therefore we observe their combined contribution at each position on the sky .
thus , it is necessary to separate their contribution to the total observed galaxy spectrum and measure the properties and kinematics of their stellar populations separately . to this aim
we developed a spectroscopic decomposition technique and successfully applied to three spiral galaxies with counter - rotating stellar disks @xcite .
we observed with the visible multi object spectrograph ( vimos ) integral - field unit at the very large telescope ( vlt ) three spiral galaxies , which are known to host counter - rotating stellar disks : ngc 3593 @xcite , ngc 4550 @xcite , and ngc 5197 @xcite .
our spectroscopic decomposition technique exploits the differences in kinematics and stellar populations of the two components to separate their contribution from the combined observed spectrum . at each position on the galaxy , the code builds two synthetic templates ( one for each stellar component ) as linear combination of spectra from an input stellar library , and convolves them with two gaussian line - of - sight velocity distributions with different kinematics .
gaussian functions are also added to the convolved synthetic templates to account for ionized - gas emission lines ( h@xmath0 , h@xmath1 , [ o iii ] , and [ n i ] ) .
the spectroscopic decomposition code returns the spectra of two best - fit stellar templates ( that represent the two stellar disks ) and ionized - gas emissions , along with the best - fitting parameters of luminosity of each component , velocity , and velocity dispersion .
the line strength of the lick indices of the two counter - rotating components are measured on the two best - fit stellar templates .
the line strength of the absorption line indices are then used to infer the age , metallicity , abundance ratios , stellar mass - to - light ratio , and the stellar mass of the two stellar components .
we successfully separated the contribution of the ionized gas and the two stellar components from the observed spectrum , and confirm the presence of two counter - rotating stellar disks in ngc 3593 .
more details on this galaxy and other galaxies in our sample are in @xcite . fig .
[ fig:003_1 ] shows the observed field of view , the 2-dimensional velocity fields of the stellar and ionized gas components , and the line strength of the absorption line indices ( h@xmath1 , mg _ b _ , [ mgfe ] , and @xmath2fe@xmath3 ) . fig .
[ fig:003_2 ] shows the 2-dimensional maps of the age , metallicity and @xmath4-enhancement , and their radial profiles .
[ fig:003_3 ] shows the radial profiles of the stellar mass - to - light ratio and the mass of the two components .
the secondary disk is on average less luminous and less massive than the main stellar disk , and it rotates along the same direction as the ionized gas component .
the two counter - rotating stellar components have different stellar populations .
the secondary stellar disk is younger , more metal poor , and more @xmath4-enhanced than the main galaxy stellar disk .
the age difference between the two components is higher within 500 pc : the main component is @xmath5 gyr old , whereas the secondary component is @xmath6 gyr old . from the difference in age of the two components , we are able to date the formation of the counter - rotating stellar disk to @xmath72 gyr ago , i.e. @xmath71.6 gyr after the formation of the main galaxy disk .
both components have high metallicity in the central 150 pc followed by a declining profile .
the main component is on average more metal rich ( @xmath8_1=-0.04\pm0.03 $ ] ) than the secondary component ( @xmath9_2=-0.15 \pm0.07 $ ] ) .
our results rule out internal processes , such as bar dissolution , since they would lead to the formation of two counter - rotating disks with the same properties .
external acquisition of gas followed by star formation , or mergers of two galaxies with similar mass are both viable mechanisms .
the accretion scenario predicts that the secondary component is the younger , and that the two components have different metallicity , as they are formed from different gas clouds . in the second scenario ,
the properties of the two stellar component depend on the properties of the merging systems and star formation episodes triggered by the merger . | we present a spectral decomposition technique and its applications to a sample of galaxies hosting large - scale counter - rotating stellar disks . our spectral decomposition technique allows to separate and measure the kinematics and the properties of the stellar populations of both the two counter - rotating disks in the observed galaxies at the same time .
our results provide new insights on the epoch and mechanism of formation of these galaxies . |
maximally entangled quantum states , when their two halves are shared between two parties , are a uniquely valuable resource for various information - processing tasks . used in conjunction with a quantum communications channel , they can increase the classical data carrying capacity of that channel , in some cases by an arbitrarily large factor@xcite .
possession of maximally entangled states can ensure perfect privacy of communication between the two parties by the use of quantum cryptography@xcite .
these states can facilitate the rapid performance of certain forms of distributed computations@xcite .
of course , maximally entangled states are the key resource in quantum teleportation@xcite . on the other hand , the surreptitious establishment of entanglement between two parties can thwart the establishment of trust between parties via bit commitment@xcite .
how can two parties come into the possession of a shared maximally entangled state ?
if the storage and transportation of quantum particles were perfect , then the state could have been synthesized in some laboratory long in the past and given to alice and bob ( our personified parties ) for storage until needed . in practice no such perfect infrastructure exists .
since the most interesting scenarios for the use of quantum entanglement are in cases where alice and bob are remote from one another , we will consider the long - distance transportation of quantum states needed to establish the shared entanglement to be difficult and imperfect , while the local processing of quantum information ( unitary transformations , measurement ) we will assume , for the sake of analysis , to be essentially perfect . under these assumptions , when we wish to assess whether a given physical setup is or is not useful for entanglement assisted information processing , our analysis focuses on the mixed quantum state , @xmath3 , in the hands of alice and bob after the difficult transportation step .
we enquire whether @xmath4 can be transformed , by _
lq+cc operations _ , to a supply of maximally entangled states . here
the @xmath5 notation indicates that @xmath6 copies of the state @xmath3 are available , and we will be concerned with asymptotic results as @xmath6 is taken to infinity .
lq+cc operations ( sometimes called locc in the literature ) are obtained by any arbitrary sequence of _ local _ quantum operations ( appending ancillae , performing unitary operations , discarding ancillae ) supplemented by _ classical communication _ between alice and bob .
an interesting fact about this possibility for the _ distillation _ of entanglement is that it is neither rare nor ubiquitous ; a finite fraction of the set of all possible bipartite mixed states @xmath3 can be successfully distilled@xcite , and a finite fraction cannot@xcite .
much work has been focussed on whether @xmath3 falls into the distillable or into the undistillable class , and this paper is primarily a contribution to this classification task . before describing our new contributions , we will give a brief review of previous results on classifying states according to their distillability .
multipartite density matrices @xmath3 are considered _ unentangled _ if there exists a decomposition of @xmath3 into an ensemble of pure product states ; for the bipartite case this means that we can write latexmath:[\[\rho=\sum_i p_i |\alpha_i\rangle\langle\alpha_i|\otimes to as separable states .
it is clear that separable states are never distillable .
however , the converse proposition , that entangled states are always distillable , is false in general , although true for density matrices in @xmath2 and @xmath8 hilbert spaces@xcite .
this became clear shortly after the introduction by peres@xcite of a computationally simple criterion for separability , the _ partial transposition _ test .
the partial - transpose operation , denoted as @xmath9 when the transpose is applied to bob s hilbert space , is specified by the action @xmath10 while application of @xmath11 to alice s hilbert space will lead to identical results , we will always apply it to bob s space in this paper . here
@xmath12 and @xmath13 indicate an orthonormal basis on alice s hilbert space , and @xmath14 and @xmath15 the same for bob .
it is easy to show that separable states are positive under partial transpose , that is , that the matrix @xmath16 is a positive semi - definite operator , denoted by @xmath17 .
( [ intronot ] ) introduces the @xmath18 notation that we will use throughout this paper . )
this positivity property is abbreviated as ppt ; states for which @xmath19 are called npt states .
it was soon recognized@xcite that the set of ppt density operators @xmath3 is larger than the set of unentangled states ( except in @xmath2 and @xmath8 ) ; see fig .
[ pptetc ] .
it was also discovered that all ppt states , even those which are inseparable , are not distillable .
the existence of such states , in which entanglement is present ( since entanglement is required to synthesize the states ) but can not be rextracted in pure form , was a surprising observation , indicating the possibility of a fundamentally new form of irreversibility in physics .
states having this property are said to possess _ bound entanglement_. the introduction of the ppt / npt classification suggested a new conjecture about distillability , namely that all states with npt would possess distillable entanglement , and it is the purpose of the present paper to explore this conjecture . while no rigorous results have been obtained concerning this conjecture , we will introduce a two - parameter family of npt states for which we obtain evidence that the conjecture is false .
that is , we consider it likely that the family of states we introduce below has only bound entanglement , despite being npt .
we have been able to recast the question about the distillability property of the @xmath0 states , or of any npt states , as a question about the two - positivity properties of certain positive linear maps@xcite .
these maps arise because there is a one - to - one correspondence between mixed states on @xmath1 and completely positive linear maps @xmath20 on @xmath21 dimensions which , when applied to half a maximally entangled state @xmath22 , produces the mixed state @xmath3 .
then , the npt property is related to the map @xmath23 and its compositions @xmath24 .
the open question about distillability can be posed compactly as a question concerning the mathematical properties of these maps .
this approach also permits us to consider the question of whether distillability is an additive property , that is , whether the amount of distillable entanglement of @xmath25 is just the sum of the two separately .
et al._@xcite have given some evidence for a kind of undistillability involving single copies of ppt bound entangled states . in the positive - map language ,
the most general questions about non - additivity can be compactly framed .
this shows that further developments of the theory of positive maps will be very desirable in settling some of the fundamental questions about the entanglement properties of quantum states .
this paper develops in the following way : section [ canon ] introduces the canonical states @xmath0 and shows the lq+cc mapping that produces them .
section [ tools ] considers the distillability of any state @xmath3 by application of the basic criterion of whether it remains entangled when projected into @xmath2 .
section [ single ] considers a single copy of the @xmath0 state , establishing the @xmath0 for which there exist such projections into @xmath2
[ multi ] takes up the much harder case of multiple copies , with sec .
[ nulls ] proving the result that for some @xmath26 states , no entanglement remains upon projection into @xmath2 even for arbitrarily large @xmath6 .
section [ twopos ] recasts the question about distillability in terms of two - positivity of linear maps isomorphic to the mixed states .
the desired , but too - ambitious , program would be to assess the distillability of all npt states .
we will attempt this assessment only for a specific subset of the npt states parameterized by two real numbers .
this subset will , however , have a specific relation to the set of npt states , in that there is a lq+cc operation that will map the general npt state onto one in our two - parameter family .
this lq+cc operation preserves the npt property .
thus , if we could exhibit a protocol for the distillation of our two - parameter family , this would suffice to show that all npt states were distillable . on the contrary , our canonical two - parameter family has properties which make distillation quite hard for certain ranges of the parameters , suggesting that in fact some portion of the full set of npt states is _ not _ distillable .
our canonical states , with real parameters @xmath27 and @xmath28 , are written as @xmath29 here @xmath30 the states live in a @xmath1 hilbert space .
the parameter @xmath31 in eq .
( [ abcform ] ) is not independent , because of the unit trace condition it is related to @xmath27 and @xmath28 by @xmath32 the range of interest for the parameters @xmath27 and @xmath28 is shown in fig .
[ nptreg ] .
as we will show in the next section , the state is npt in two triangular regions of parameter space ; one of these regions @xmath33 , which will not be of much interest to us ( all these states are distillable ) , lies above the straight line @xmath34 , and is defined by the inequality @xmath35 .
the region @xmath36 , about which we will have much more to say , lies in the region @xmath37 and is defined by @xmath38 .
region @xmath39 contains ppt states ; in sec .
[ multi ] we prove that all these states are also separable . to show that @xmath0 represents a canonical set
, we will exhibit a procedure involving only lq+cc operations that will convert any npt density matrix @xmath3 , that is , one satisfying the condition @xmath40 for some state @xmath41 , to one of the @xmath0 form having npt .
we will take the hilbert space dimension to be @xmath42 , that is , we will not restrict alice s and bob s dimensions to be the same . here is the sequence of lq+cc operations that will reduce the general npt state @xmath3 to @xmath0 : \(i ) _ rotation to the schmidt basis : _ we write the @xmath41 of eq .
( [ negfirst ] ) as @xmath43 here @xmath44 .
let @xmath45 and @xmath46 , or @xmath47 where @xmath48 we define @xmath49 .
equation ( [ negfirst ] ) can be rewritten as @xmath50 where @xmath51 is transposition in a rotated basis determined by @xmath52 .
the negativity of the expression eq .
( [ negin ] ) does not depend on the basis in which @xmath11 is performed , therefore we will replace @xmath51 by @xmath11 again in the remainder .
\(ii ) _ local filtering ( see @xcite ) : _ we define the state @xmath53 as @xmath54 the filter operation @xmath55 on alice s hilbert space is defined by the equation @xmath56 we apply this local filter to the state @xmath57 to obtain @xmath58 : @xmath59 eq .
( [ negfirst ] ) implies that @xmath60 we now use that @xmath61 and @xmath62 to rewrite this npt condition in a form which will be convenient below : @xmath63 with @xmath64 this hermitian operator @xmath65 can be written in its eigenbasis : @xmath66 where @xmath30 \(iii ) _ project into @xmath1 : _ since @xmath53 , and @xmath65 , have support only a @xmath1 dimensional subspace of the hilbert space , alice and bob can project locally onto this subspace and leave the npt condition eq .
( [ nptc1 ] ) , or eq . ( [ nptc2 ] ) , unchanged .
we call the resulting npt density matrix in @xmath1 @xmath67 .
\(iv ) _ diagonal twirl : _ alice and bob perform a equal mixture of identical unitary operations , which are diagonal in the schmidt basis given by the vectors @xmath12 , giving state @xmath68 .
this unitary operation is @xmath69 the phases @xmath70 are chosen randomly over a uniform distribution from 0 to @xmath71 , independently for each @xmath72 .
this leaves the operators @xmath73 invariant .
this operation therefore leaves the eigenvectors of @xmath65 and thus @xmath65 itself invariant .
thus it follows that @xmath74 the ` twirled ' density matrix @xmath68 has the form : @xmath75 note that the coefficients in these sums are all in general distinct , with @xmath76 not necessarily equal to @xmath77 and similarly for @xmath78 .
\(v ) _ symmetrize by permutation : _ alice and bob carry out identical , randomly chosen unitary transformations which are drawn uniformly from all possible permutation operations over the elements of the schmidt basis @xmath12 .
this ensures that in the new density matrix @xmath79 the @xmath80 coefficients , for all @xmath72 , become equal to a single number @xmath31 , all the @xmath76 become equal ( we call this constant @xmath81 ) , and all the @xmath78 become equal ( we call this constant @xmath82 ) .
so we obtain @xmath83 but comparing with eq .
( [ abcform ] ) , we note that we have arrived at the desired canonical form , @xmath84 as the hermitian matrix @xmath65 of eqs .
( [ nptc2 ] ) and ( [ nptc3 ] ) is again invariant under this symmetrization , we note that the npt property is again preserved : @xmath85 we may summarize the foregoing line of argument as a theorem : let @xmath3 be a bipartite density matrix on @xmath86 with the property that @xmath19 .
the density matrix @xmath3 can be converted by local operations and classical communication to a density matrix @xmath0 on @xmath1 with @xmath87 characterized by two real parameters @xmath27 and @xmath28 such that @xmath88 .
this density matrix @xmath0 is @xmath89 with @xmath90 it is easy to see from the form of @xmath65 that these transformations carry all npt states @xmath3 into a @xmath0 sitting in the @xmath36 region of fig .
[ nptreg ] .
this is why the @xmath33 region will not be of concern to us .
we note that it is possible to follow the five - step reduction above with another lq+cc operation , resulting in a canonical npt density operator characterized by just a single real parameter : \(vi ) _ full twirl : _ alice and bob perform a equal mixture of identical unitary operations drawn uniformly ( with the haar measure ) from the entire group @xmath91 .
it is straightforward to show that the resulting density matrix @xmath92 has the same form as above ( eq . ( [ abcform2 ] ) ) : @xmath93 with @xmath94 and @xmath95 given by the same constraint as in eq .
( [ constr ] ) .
thus , @xmath92 depends only on the single parameter @xmath27 ; it is the same werner density matrix studied recently by horodecki _
et al._@xcite : @xmath96 , \label{werner}\ ] ] note that @xmath65 of eq .
( [ nptc2 ] ) is proportional to the `` swap '' operator @xmath97 this `` full twirl '' carries all the states in the @xmath37 region of fig .
[ nptreg ] onto the line @xmath98 , without changing the value of @xmath27 .
of course , if it were possible to prove that all the npt states of the one - parameter form @xmath99 were distillable , then all npt states would be distillable through the reductions we have developed above . in fact
we conjecture , as horodecki _
et al . _ have previously ( sec .
viii , ref .
@xcite ) , that some of these npt states are undistillable . under these circumstances , it is desirable to provide evidence for undistillability for the widest class of states possible , and we will concentrate in this paper on providing such evidence for the two - parameter family of canonical states @xmath0 , more particularly , for those lying near the line segment @xmath100 in fig .
[ nptreg ] .
all of the results we develop will , of course , also apply to the restricted one - parameter family @xmath99 as well .
in this section we will explore all the known tools at our disposal for analyzing the distillability of states . for some of the @xmath0 states we believe that no distillation protocol exists ; evidence for this
is provided by the last result of this section , that for some @xmath0 states , any successful distillation protocol , if it exists , must act on some very large number @xmath6 of copies of the state ; we show that @xmath6 must diverge along an entire boundary @xmath100 in fig .
[ nptreg ] . much of the discussion of distillation strategies will need the notion of the _ schmidt rank _ of a pure state in an ensemble decomposition of density matrix @xmath3 .
we first define this term : a bipartite pure state @xmath41 has schmidt rank k if the state can be written in the schmidt polar form as @xmath101 with @xmath102 and @xmath103
. the distillation of the @xmath0 states ( or more particularly , of the @xmath99 subset of these states ) has already been considered in@xcite . there , a distillation protocol was developed based on the positive linear map @xmath104 . in sec .
[ twopos ] we will discuss other aspects of the relation between the theory of positive maps and the distillability of mixed states . for all states @xmath3 for which @xmath105 , it was shown how to distill them by converting these states to a different canonical density - matrix form introduced by werner .
however , all the states @xmath0 remain positive under the action of @xmath106 , so long as the dimension @xmath107 , because @xmath108 , where @xmath109 and @xmath110 .
( positivity under the action of @xmath106 was already known for @xmath99@xcite . )
thus , the simple distillation procedure studied in@xcite will not work for these states .
thus , to study the distillability of these states , we need to consider the more general necessary and sufficient condition developed by horodecki _
: ( horodecki et al.@xcite ) a density matrix @xmath111 is distillable if and only if there exists a finite @xmath6 and projections @xmath112 and @xmath113 such that @xmath114 is entangled .
[ horlem ] in @xmath2 , a density matrix @xmath115 is entangled if and only if it is npt .
lemma [ horlem ] requires the examinations of projection of the density matrix ( or @xmath6 copies of the density matrix ) .
the following lemma gives a convenient recasting of these properties of projections in terms of properties of the original density matrix itself : [ x2n ] let @xmath3 be a density matrix on @xmath116 .
let @xmath117 be a projection and also @xmath118 .
there exist @xmath119 and @xmath120 such that @xmath121 is entangled if and only if @xmath122 has the property that @xmath123 eq . ( [ notpsd ] ) is equivalent to the condition that there exists a state @xmath124 that has schmidt rank two and @xmath125 _ proof : _ if the density matrix @xmath126 is not positive semidefinite under partial transposition , then there exists a schmidt rank two vector @xmath41 , written in its schmidt basis as @xmath127 such that @xmath128 ( the state @xmath41 can not be a product vector since , if it were , @xmath129 . ) we note that the projector @xmath119 in eq . ( [ oneside ] )
consistent with eq .
( [ sd ] ) has the form @xmath130 .
note also that the state @xmath41 is invariant under the projector @xmath131 , @xmath132 plugging eqs ( [ inv ] ) and ( [ oneside ] ) into eq .
( [ 2nneg ] ) : @xmath133^{pt}({\bf 1}\otimes p^\dagger_b){| \psi \rangle}= { \langle \psi | } [ ( p_a \otimes p^*_b)\ , \rho\ , ( p^\dagger_a \otimes p^t_b)]^{pt } { | \psi \rangle } < 0.\label{long1}\ ] ] therefore the state @xmath134 on @xmath2 is entangled .
conversely , if the density matrix @xmath126 is positive semidefinite under partial transposition for all @xmath119 , meaning that @xmath126 is either separable or has bound entanglement , then there does not exist a @xmath120 such that @xmath135 is entangled , because then it could be distilled . finally , by rewriting eq .
( [ long1 ] ) as @xmath136 we note that @xmath137 is the state needed for eq .
( [ negexp ] ) .
@xmath138 note that an easy consequence of lemma [ x2n ] is that all npt states in @xmath139 for any @xmath6 are distillable .
the real difficulty in applying lemma [ horlem ] is that it requires the examination of an arbitrary number of copies @xmath6 of the state to be distilled .
we will therefore first develop a set of strong results for the special case of @xmath140 , then we will move on to obtain some results for the much more difficult case of arbitrary @xmath6 .
we begin with some terminology : : we say that density matrix @xmath3 is _ pseudo one - copy undistillable _ if , for all schmidt rank two states @xmath124 , @xmath141 .
then , by lemma [ x2n ] , there exists no @xmath142 projection of @xmath3 that is inseparable .
we say @xmath3 is _
pseudo @xmath6-copy undistillable _ if and only if @xmath4 is pseudo one - copy undistillable
. we will establish which states @xmath0 are pseudo one - copy undistillable and which are distillable .
the partial transpose of @xmath0 reads @xmath143 the eigendecomposition of @xmath144 is @xmath145 with @xmath146 which we refer to as the `` e - dit eigenstates '' in analogy with `` ebit '' , because they are the maximally entangled states in @xmath147 having a `` dit''(@xmath148 bits ) of entanglement . correspondingly , we refer to the @xmath149 states with @xmath150 as the `` product eigenstates '' .
the eigenvalues @xmath151 are given by @xmath152 the negative eigenvalue @xmath153 is independent of @xmath28 , showing why the ppt - npt boundary is a vertical line ( @xmath100 in fig [ nptreg ] ) .
notice that the eigenvectors of @xmath144 are independent of parameters @xmath27 and @xmath28 .
we now specialize to the state for which the positive eigenvalues are all equal , @xmath154 , and therefore @xmath155 these are precisely the werner states @xmath99 of eq .
( [ werner ] ) above , the states along the line @xmath98 in fig .
[ nptreg ] .
we take advantage of the fact that lemma [ x2n ] does not require normalized states to write the partial transpose of these states in the following simple unnormalized form : @xmath156 with @xmath157 .
we will show that for @xmath158 , @xmath159 and that for @xmath160 , @xmath161 , with the minimum taken over all schmidt rank two vectors .
thus @xmath162 , corresponding to @xmath163 and @xmath164 ( the point @xmath165 in fig .
[ nptreg ] ) is the transition point separating distillable werner states ( line segment @xmath166 ) from those which are pseudo one - copy undistillable ( line segment @xmath167 ) . to establish this we first need to prove the following lemma : [ lem : overlapschmidt ] in @xmath1 ,
the overlap of a schmidt rank two state with a maximally entangled state is at most @xmath168 . in other words ,
if @xmath169 has schmidt rank two and @xmath170 is a maximally entangled state , then @xmath171 _ proof : _ in its schmidt basis , @xmath172 .
since @xmath169 is schmidt rank two , it may be written in its schmidt decomposition as @xmath173 , with @xmath174 .
the overlap then is , @xmath175 where @xmath176 is the vector obtained by complex conjugation of the components of @xmath177 in the schmidt basis of the state @xmath170 .
thus , we have @xmath178 maximizing with constraint @xmath174 gives the desired result .
@xmath138 now we are ready for the main result : [ 1shot ] given @xmath179 whose partial transpose is given in eq .
( [ eq : sigmapt ] ) , we have , * if @xmath158 then @xmath115 is not pseudo one - copy distillable . * if @xmath160 then @xmath115 is pseudo one - copy distillable . _ proof : _ we start with the first part .
let @xmath169 be any schmidt rank two vector .
then , @xmath180 where we have used lemma [ lem : overlapschmidt ] .
this is greater than or equal to zero for @xmath158 , showing the first part of the result .
for the second part , consider @xmath181 .
we have @xmath182 , which is less than zero for @xmath160 , proving the second part of the result .
@xmath138 from this it is a simple matter to completely characterize the one - copy undistillability of the @xmath0 states : [ abc1shotno ] the states @xmath0 are pseudo one - copy undistillable in the region of parameter space bcgk in fig .
[ nptreg].@xcite _ proof : _ since any state in the region is a convex linear combination of the states @xmath183 , @xmath184 , @xmath165 , and @xmath185 , it suffices to show that the partial transpose of each of these four states has a positive expectation value with respect to any schmidt rank two vector ( lemma [ x2n ] ) .
this is obviously true for the ppt states @xmath183 and @xmath185 , and it is true for state @xmath165 by theorem [ 1shot ] . to show it for @xmath184 , which has parameters @xmath186 , @xmath187
, we note that the partial transpose of the state @xmath184 can be written @xmath188 here @xmath189 is the partial transpose of the normalized state at point @xmath165 , and @xmath190 is the normalized projector @xmath191 . the expectation value of the first term on the right - hand side of eq .
( [ deco ] ) is positive by theorem [ 1shot ] , and that of the second term is positive because it is a projector.@xmath138 all other states are distillable : [ abc1shot ] the states @xmath0 are distillable in the region of parameter space cfkg in fig .
[ nptreg].@xcite _ proof : _ in the region @xmath192 the partial transpose has a negative expectation value with respect to the schmidt rank two state @xmath193 , and in the region @xmath194 with respect to the state @xmath195 it has proved to be much harder to obtain definitive results concerning the pseudo @xmath6-copy undistillability of the @xmath0 states .
but we have accumulated various pieces of evidence , which we will present here , all indicating the likelihood that many of the @xmath196 states are undistillable .
our attention will focus here on a particular subset of the @xmath0 states labeled by @xmath28 and a small parameter @xmath197 , which sit just to the right of the line segment @xmath100 in fig .
[ nptreg ] : @xmath198 the eigenvectors of the partial transpose of this state @xmath199 are given in eqs .
( [ eigvecs1]-[eigvecs2 ] ) , since these are common to all @xmath0 states .
the eigenvalues are @xmath200 , @xmath201 and @xmath202 .
the only properties of these eigenvalues that we will use is that for small , positive @xmath197 and @xmath203 , @xmath153 is negative and goes to zero as @xmath204 , and @xmath205 and @xmath206 are strictly positive .
although we will not need any more properties of the density matrices @xmath207 , we can at this point note the interesting fact that they are all separable ; in fact , _ all _ the ppt states of the form @xmath0 ( the region @xmath39 in fig . [ nptreg ] ) are separable ( eq . ( [ defsep ] ) ) .
this is established by showing that the density matrices at the extremal points @xmath208 , @xmath183 , @xmath185 , and @xmath209 are separable ; all other states in this region are convex combinations of these .
the state at @xmath208 is proportional to @xmath210 , and the one at @xmath185 is proportional to @xmath211 , so these are both obviously separable . we can also create the state @xmath212 at point @xmath183 using separable states .
it is easiest to construct this ensemble for the partial transpose of this state ( see eq .
( [ abcpt ] ) ) , which is done by equally mixing the states @xmath213 for all pairs @xmath214 , and @xmath215 . by mixing these states with equal probabilities , all terms of the form @xmath216 , @xmath217 and @xmath218 for @xmath219 cancel out ;
each of these will come with a factor @xmath220 or @xmath221 . a term such as @xmath222 will occur @xmath223 times as much as a term @xmath224 , which is indeed the correct ratio for @xmath225 .
the state @xmath212 itself at point @xmath183 is obtained from mixing the states @xmath226 with equal probabilities .
the partial transpose of the state at point @xmath209 has a simple form ( @xmath227 in eq .
( [ eigvecs1 ] ) ) ; it is straightforward to show that @xmath228 at @xmath209 is realized by an equal mixture of the separable states @xmath229 where each integer @xmath230 , @xmath231 , ... @xmath232 runs independently over 0 , 1 , and 2 .
this is clearly not a separable decomposition with the minimal possible number of states .
a few notes about the decomposition for point @xmath183 : for @xmath233 the state @xmath228 at point @xmath183 has rank eight .
this implies that the optimal decomposition of @xmath228 , and therefore of @xmath3 itself , needs at least eight states in its decomposition ; this despite the fact that the rank of @xmath3 is only six ( see lemma 1 of ref .
thus we have a new example of a state for which the number of states in its minimal decomposition exceeds its rank ; but see ref .
for general @xmath21 , the number of states in our separable ensemble at @xmath183 , @xmath234 , which is more than the dimension @xmath235 for @xmath236 .
there are no known prior explicit examples in which the number of members of the optimal ensemble is greater than the dimension ; it would be interesting to prove that eq .
( [ dec1 ] ) constitutes a minimal optimal ensemble .
the separability of the ppt states permits us to give an extension of proposition [ abc1shotno ] indicating that the undistillability of states in this region is linked : [ momo2 ] if the state @xmath0 at point g is pseudo n - copy undistillable , then all states in the region bgk are pseudo n - copy undistillable . _
proof : _ first , note that if the state at point @xmath165 is pseudo @xmath6-copy undistillable , then it is also pseudo @xmath237-copy undistillable for @xmath238 . since the two extremal points @xmath183 and @xmath185 of the convex set of states @xmath239 are separable , the partial transpose of all states in this region can be written as a convex combination ( using notation from eq .
( [ deco ] ) ) : @xmath240 where @xmath241 are product projectors and @xmath242 . applying lemma [ x2n ]
, we consider the expectation value of @xmath6 copies of this state with respect to any schmidt rank two vector @xmath169@xcite : @xmath243 we need to show that this is non - negative ; we show this by demonstrating that each term in the tensor product , when expanded out , is not negative . consider a term containing @xmath237 @xmath189 factors and @xmath244 factors involving the projectors @xmath241 .
we can apply the @xmath244 projectors to @xmath169 ; since they are all product projectors , the projected vector @xmath245 still has schmidt rank two ( or one ) .
so , the matrix element of eq .
( [ bigme ] ) is proportional to @xmath246 but if @xmath165 is pseudo @xmath237 copy undistillable , this matrix element is non - negative.@xmath138 note that this analysis does not apply to state @xmath184 , because the projectors @xmath190 of eq .
( [ deco ] ) are not product projectors ; therefore , they can increase the schmidt rank of @xmath169 . for @xmath233
we have performed extensive numerical studies to search for states distillable by projection on two copies in the region @xmath247 .
we find none , reinforcing the indication of lemma [ momo2 ] that an entire region inside the @xmath196 set will prove to be undistillable
. the next section will provide further evidence for this idea .
in this section we will obtain our strongest result , which suggests that some of the npt states @xmath0 are not distillable .
we will be able to conclude that for any finite @xmath6 there exists an @xmath197 such that @xmath248 ( eq . ( [ epsstates ] ) ) is not entangled on any @xmath249 subspace , and is therefore one - copy undistillable .
this result can have only one of two further implications : 1 ) for some @xmath28 , this @xmath197 asymptotes to some finite value @xmath250 as @xmath251 . in this case ,
the npt states @xmath252 are absolutely undistillable .
2 ) for all @xmath28 , this @xmath197 goes to zero as @xmath251 . in this case
all states immediately to the right of line @xmath100 are distillable ; thus all @xmath0 states with npt would be distillable , since all such states can be first mixed with some separable @xmath0 state ( a lq+cc operation ) to bring it to the @xmath100 line .
but , one might say that the states near @xmath100 are `` barely '' distillable : an arbitrarily large number of copies of the state are required before there is any sign of undistillability of the state .
it would be fair to say that these states would still be undistillable in any practical sense .
first , we establish the significance of the null - space properties of @xmath207 for the argument .
we consider the function @xmath253 here the minimum is taken over all schmidt rank two states @xmath254 in the full @xmath255 hilbert space . by lemmas [ horlem ] and [ x2n ] , we know that the sign of @xmath256 determines whether @xmath257 is pseudo n - copy undistillable
. for @xmath258 the state is separable and therefore @xmath259 for all @xmath6 .
the question is , does the state become pseudo n - copy undistilable as @xmath260 ?
the answer is provided by the result whose proof we outline in a moment , that there is no schmidt rank two vector in the null space of @xmath261 for any @xmath6 . in other words , for all @xmath6 and @xmath28 , @xmath262 and , since @xmath263 is a continuous function of @xmath197 , there must therefore exist an @xmath264 such that @xmath265 thus , there is a finite range of @xmath266 for every @xmath6 ( and for all @xmath28 such that @xmath203 ) such that @xmath257 is pseudo n - copy undistillable .
the only knowledge lacking at this point for a complete demonstration of the undistillability of @xmath257 is the asymptotic behavior of @xmath267 as @xmath251 .
if @xmath268 as @xmath251 , then npt undistillability would not be established ; we would merely have shown that distillation becomes difficult as @xmath260 , requiring more and more copies of the state in the distillation protocol . if @xmath267 remains larger than some positive @xmath250 for all @xmath6 and for some @xmath28 , then we would know that all states @xmath269 are absolutely undistillable . since the signs from our few - copy work are that indeed this threshold remains positive we are led to the conjecture : * conjecture : * _ states @xmath257 of eq .
( [ epsstates ] ) , for sufficiently small positive @xmath197 , are undistillable .
_ we can further speculate that the undistillable region will correspond exactly to region @xmath247 in which the state is pseudo one - copy and , apparently , pseudo two - copy undistillable .
it may well be that pseudo one - copy undistillability and absolute undistillability are equivalent .
now we present our result about the null - space properties of @xmath207 on which the above discussion is based : its null space does not contain any non - zero vectors of schmidt rank less than three .
first we set up some notation .
plain roman indices take values from @xmath270 to @xmath223 unless otherwise stated .
let indices with superscript @xmath271 represent composite indices , e.g. @xmath272 represents @xmath273 .
label the eigenvectors ( eq .
( [ eigvecs1 ] ) ) as @xmath274 and @xmath275 with @xmath276 . the label @xmath277 stands for `` e - dit eigenstate '' and the label @xmath271 stands for `` product eigenstate '' .
let us denote @xmath6-tuples of indices such as @xmath278 by letters in bold font such as * i * ; in sums over * i * , each @xmath279 runs independently between @xmath270 and @xmath223 .
next we prove an important lemma : [ lem : first ] the null space of the partial transpose of density matrix @xmath280 , for all @xmath28 , @xmath281 and @xmath282 , does not contain any non - zero vectors with schmidt rank less than three of the form @xmath283 _ proof : _ for @xmath140 the result is obvious since @xmath284 is the only vector in the null space and it has schmidt rank @xmath281 . for @xmath285
, we first note that the partial trace of the state in eq .
( [ bigvec ] ) is @xmath286 where the coefficients @xmath287 are the @xmath6-dimensional discrete fourier transforms of the @xmath31 s . here
note that @xmath289 is already diagonalized .
since the schmidt rank of a pure state equals the rank of the partial trace , we require that the rank of @xmath289 be less than three .
thus at most two coefficients @xmath290 are nonzero , i.e. @xmath291 where @xmath292 . solving for the @xmath293s by doing an inverse fourier
transform we have @xmath294 and @xmath295 must be nonzero and @xmath296 .
now we start putting constraints on the @xmath31 s such that the vector @xmath297 is in the null space of @xmath298 .
the vector @xmath299 belongs to the null space only if @xmath300 , because the corresponding eigenvalue @xmath301 is positive at @xmath302 .
now we impose the null space constraint @xmath303 ( since the corresponding eigenvalue @xmath304 is positive at @xmath302 ) , and we have @xmath305 .
similarly , other @xmath31 s , whose subscripts are obtained by permuting @xmath306 , may be constrained to zero giving @xmath307 .
however this implies that the vector is of schmidt rank one if it is to satisfy these null space constraints .
thus no schmidt rank two vector of the form @xmath308 belongs to the null space of @xmath298 .
next we consider the case of schmidt rank one vectors , where without loss of generality we may assume @xmath309 .
then , the null space constraint @xmath300 implies that @xmath310 , thus proving the result .
@xmath138 now we are ready for the main result : the null space of @xmath311 for @xmath312 and @xmath282 does not contain any vector of schmidt rank less than three . _
proof : _ for @xmath140 the result is obvious , because the null space consists of the span of the vector @xmath313 which has schmidt rank @xmath312 . for purpose of illustrating the proof technique , we next prove the result for two copies , i.e. , @xmath314
. then we will show how the proof generalizes to @xmath6 copies .
recalling eq .
( [ eigvecs1 ] ) and the fact that the eigenvectors form a basis for the one - copy hilbert space of @xmath315 , a general vector @xmath41 in the hilbert space of two copies can be written as @xmath316 with @xmath317 , @xmath318 , @xmath319 and @xmath320 . here
the @xmath321 s are complex coefficients for the vectors
. the @xmath322 term must be zero if the vector is to belong to the null space , because the corresponding eigenvalue @xmath323 is positive at @xmath302 . now assuming @xmath324 has schmidt rank less than three
, we will show that the coefficients @xmath325 s and @xmath326 s are zero . to show this we repeatedly use the fact that local projections can not increase the schmidt rank of a vector .
alice and bob can project locally on the vector @xmath327 , for any @xmath328 of the first copy , which results in a vector proportional to @xmath329 .
by lemma [ lem : first ] this vector has schmidt rank greater than two unless it is zero .
thus all the @xmath326 s are zero . similarly applying this argument to the @xmath330 term , with the projection now done on a product vector @xmath327 of the second copy
, we see that the @xmath325 s are zero .
the only term left now is the @xmath331 term , for which lemma [ lem : first ] applies and gives us the result .
we write the general proof for @xmath6 copies along the lines of the two - copy proof , albeit with considerable notational complications . generalizing the notation of eq .
( [ twocop ] ) , we define @xmath332 to be the set of all distinct permutations of @xmath237 @xmath271 s and @xmath333 @xmath277 s .
we also denote the strings representing permutations in @xmath334 by bold font , e.g. , @xmath335 , where the @xmath336 are the characters in the permutation string , e.g. , for @xmath337 , then @xmath338 , @xmath339 , and @xmath340 .
a general state in the @xmath6-copy hilbert space can be written in the form @xmath341 with @xmath342 again , the @xmath343 term is zero if the vector is to be in the null space , because the corresponding eigenvalue @xmath344 is positive at @xmath302 .
define @xmath345 by @xmath346 then to prove the result we show that there is no vector with schmidt rank less than three of the form @xmath347 for all @xmath348 .
this we show by induction on @xmath349 . for @xmath350
, the result immediately follows from lemma [ lem : first ] . for the induction step , we write @xmath351 now
if alice and bob locally project @xmath352 onto @xmath353 of the @xmath72th copy , for @xmath354 , the result is @xmath355 since local projection can not increase the schmidt rank , by lemma [ lem : first ] the vector inside the sum above must be zero . doing this for all the different values of the @xmath356 s we see that @xmath357 , where the superscript contains @xmath349 @xmath271 s and @xmath358 @xmath277 s .
similarly we can prove that @xmath359 is zero for any permutation string @xmath360 .
this shows that @xmath352 has to be of the form @xmath361 , for which the result is true by the induction hypothesis .
in this section we find a formulation of the problem of distillability of an arbitrary bipartite density matrix @xmath3 .
this formulation uses the notion of @xmath362-positive linear maps
. we will explicitly show how the problem of distillability of the density matrices @xmath0 that were discussed in the preceding sections can be cast in the language of positive linear maps .
let us first recall the definition of a @xmath237-positive linear map @xcite .
let @xmath363 denote the matrix algebra of operators on an @xmath6-dimensional hilbert space and let @xmath364 denote the set of positive semidefinite matrices .
a linear map @xmath365 is called positive when @xmath366 , that is , the map preserves the set of positive semidefinite matrices .
a linear map @xmath367 is called @xmath237-positive when the map @xmath368 is positive .
note that @xmath369-positivity is equivalent to positivity .
it is not hard to show that when a map @xmath370 is @xmath6-positive , it is completely positive
. we will now give an alternative characterization of @xmath237-positivity .
the next lemma says that to test a linear map for @xmath237-positivity we only need to apply it to pure states of at most schmidt rank @xmath237 .
a positive linear map @xmath371 is @xmath237-positive if and only if @xmath372 for all vectors @xmath373 which have schmidt rank at most @xmath237 .
[ kpos ] _ proof : _ if eq .
( [ kposeq ] ) holds for all states @xmath41 of schmidt rank at most @xmath237 , then it follows that @xmath374 for all vectors @xmath375 .
therefore @xmath376 for all @xmath377 and thus @xmath378 is @xmath237-positive . on the other hand ,
if there exists a vector @xmath41 of at most schmidt rank @xmath237 for which @xmath379 , then @xmath378 can not be @xmath237-positive .
@xmath138 we would like to make an additional simplification in characterizing @xmath362-positive maps .
the next lemma says that in order to test a linear map for @xmath362-positivity we only need to apply it to maximally entangled pure states of schmidt rank two .
a linear positive map @xmath371 is @xmath362-positive if and only if , for all @xmath380 with @xmath381 , @xmath382 , @xmath383 [ prop2pos ] the proof of this lemma is given in appendix [ prooflem ] .
it is possible to formulate a similar lemma for @xmath237-positive maps , in which @xmath237-positivity or the lack thereof can be deduced from applying the map on all maximally entangled vectors of schmidt rank @xmath237 . with a hermitian operator @xmath384 we can always associate a hermiticity - preserving linear map @xmath378 in the following way : @xmath385 where @xmath386 in the appendix of @xcite it was proved that the operator @xmath65 is positive semidefinite if and only the linear map @xmath378 is completely positive . from this
we conclude that any bipartite density matrix @xmath3 on @xmath315 can always be written as @xmath387 where @xmath388 is a completely positive map .
note that @xmath20 need not be trace preserving . as an example , we derive the completely positive map @xmath389 associated with the density matrices @xmath0 given in eq .
( [ abcform ] ) .
we can specify @xmath389 on the input states : @xmath390 the following main theorem expresses the connection between @xmath362-positivity and distillability of a density matrix @xmath3 : let @xmath3 be a bipartite density matrix on @xmath315 .
let @xmath391 be a completely positive map which is uniquely determined by @xmath392 let @xmath393 be a linear positive map defined as @xmath394 where @xmath11 is matrix transposition in the basis @xmath395 .
there exists no projections @xmath396 and @xmath397 such that @xmath398 is entangled if and only if the map @xmath378 is 2-positive .
let @xmath399 the density matrix @xmath3 is not distillable if and only if for all @xmath400 the map @xmath401 is 2-positive .
[ 2posing ] _ proof : _ we will prove the theorem in two parts .
first we will prove the relation between @xmath362-positivity of @xmath378 and the nonexistence of a @xmath249 subspace on which @xmath3 is entangled .
then we prove the result relating undistillability to @xmath362-positivity of @xmath401 .
let us assume that there does not exist a @xmath249 subspace on which the density matrix @xmath3 is entangled .
we can write any projector @xmath402 as @xmath403 where @xmath404 .
lemma [ x2n ] implies that @xmath405 \geq 0 , \label{ptpa}\ ] ] for all projectors @xmath119 .
this expression , using the eqs .
( [ rhoins ] ) and ( [ defl ] ) , is equal to @xmath406 with @xmath407 defined as @xmath408 the vectors @xmath409 are defined as @xmath410 . note that @xmath411 .
we now invoke the property of a 2-positive map as given in lemma [ prop2pos ] ; if eq .
( [ shouldbepos ] ) holds for all @xmath412 with @xmath413 , then @xmath378 is a 2-positive map .
conversely , invoking lemma [ prop2pos ] , if @xmath378 is a 2-positive linear map , then eq .
( [ shouldbepos ] ) holds for all states @xmath414 .
this implies that eq .
( [ ptpa ] ) holds for all projectors @xmath119 and thus there does not exist a @xmath415 subspace on which @xmath3 is entangled .
now we turn to the second part of the proof .
the necessary and sufficient condition for distillability of a density matrix was given in lemma [ horlem ] .
let @xmath416 on @xmath417 .
the density matrix @xmath3 is undistillable if and only if there exists no projections @xmath418 and @xmath419 such that @xmath420 is entangled .
thus if a density matrix is undistillable , we have , similar as eq .
( [ ptpa ] ) , @xmath421 \geq 0 , \label{ptpak}\ ] ] for all projectors @xmath422 and all @xmath400 .
we use the fact that @xmath423 is equivalent ( up to a unitary transformation ) to @xmath424 where @xmath425 is matrix transposition in @xmath426 .
( [ ptpak ] ) can be rewritten as @xmath427 for all maximally entangled states @xmath428 for all @xmath400 .
this implies with lemma [ prop2pos ] that @xmath401 is @xmath362-positive for all @xmath400 .
conversely , when @xmath401 is not @xmath362-positive for some @xmath6 , there will exist a @xmath249 subspace on which @xmath4 is entangled .
@xmath138 _ remarks : _ note that the theorem also holds for entangled density matrices @xmath3 that have the ppt property or density matrices which are separable .
in this case , however , the positive map @xmath378 is completely positive , and therefore the map @xmath401 for all @xmath400 is @xmath362-positive trivially .
we note that theorem [ 2posing ] can also be made to apply to a situation in which one is given a large number of copies of , say , two different density matrices @xmath429 and @xmath430 . with each of these density matrices
we associate a positive linear map @xmath431 and @xmath432 .
distillability of @xmath429 and @xmath430 together can be formulated as the problem of determining whether @xmath433 is @xmath362-positive .
this provides a method for searching for nonadditivity in the property of distillability@xcite .
we could encounter a situation in which both @xmath429 and @xmath430 are undistillable , but @xmath429 and @xmath430 taken together are distillable .
in general , given two @xmath362-positive maps @xmath431 and @xmath432 , the tensor product @xmath434 is not necessarily @xmath362-positive . as an example we take @xmath431 to be the identity map @xmath435 and @xmath432 a @xmath362-positive map which is not @xmath436-positive . then by definition , @xmath437 is not positive . in the cases that we consider here however , the maps are of a special form , namely @xmath438 , where @xmath20 is completely positive . for this special form , it is possible that the composed maps are always 2-positive .
the positive map @xmath439 of eq .
( [ compos ] ) corresponding to the example eq .
( [ sexamp ] ) is @xmath440 for the states on the line @xmath98 this corresponds to the positive map @xmath441 which acts as @xmath442 where @xmath443 is the parameter in eq .
( [ ldef ] )
. it has been shown @xcite that this map @xmath441 is 2-positive in the region @xmath444 .
this thus establishes an alternative proof of theorem [ 1shot ] in section [ single ] .
our alternative formulation of the problem of distillability in terms of the @xmath362-positivity property of linear maps has not yet led to a solution of the problem of npt density matrices which are ( likely to be ) undistillable ( conjecture at the end of sec .
[ nulls ] ) .
we present the formulation here , as it points to a new connection between the structure of positive linear maps and the classification of bipartite mixed state entanglement .
we expect that fruitful results will flow from understanding in more detail the classification schemes for these npt states that are based directly on their 2-positivity properties . in conclusion
, we have shown that most of the distillability properties of npt mixed states can be restricted to the study of the canonical set @xmath0 .
many of the questions about one - copy and few - copy distillability of these states are completely answered by our analysis .
a final , general proof of the full undistillability of these states eludes us , but we have shown that if they are distillable , it involves a much more difficult protocol than any which has been needed up until now . _
note added : _ after the completion of the calculations reported here , we became aware of closely related work by dr _
. this paper studies the states along the line _ hgf _ in fig .
[ nptreg ] ; for these states it provides an alternative proof to the one discussed here in sec .
[ nulls ] that , approaching point @xmath65 , the states are pseudo @xmath6-copy undistillable for any @xmath6 .
@xcite also obtains the same theorem as here ( our theorem [ 1shot ] ) about pseudo one - copy distillability of these states , as well as obtaining additional numerical results indicating that the region of pseudo two- and three - copy undistillability is the same as that for one - copy undistillability .
all the results of ref . @xcite and the present work are consistent .
dpd , jas , and bmt acknowledge support from the army research office under contract number daag55 - 98-c-0041 .
avt acknowledges support from the army research office under contract number daag55 - 98 - 1 - 0366 .
pws is supported in part by darpa through caltech s quantum information and computation ( quic ) project administered by the army research office under grant no .
daah04 - 96 - 1 - 0386 .
we are grateful to r. jozsa and m. jeng for permission to use several of their results in this paper .
we thank charles bennett , micha horodecki , pawe horodecki , and armin uhlmann for interesting discussions .
the proof is similar in structure to the proof of the lemma in the appendix of @xcite . by definition
a linear positive map @xmath445 is 2-positive if and only if , for _ all _ @xmath446 , @xmath447 we will show that we only need to consider states @xmath41 that are maximally entangled .
note that any ( unnormalized ) maximally entangled state can be written as @xmath380 with @xmath448 , @xmath449 .
we start with the following observation : when we apply the map @xmath378 on some maximally entangled state in @xmath450 , @xmath451 the matrix @xmath452 uniquely determines the action of the map @xmath378 on any input matrix that has support on the two dimensional space spanned by the vectors @xmath453 and @xmath454 . for
the first part of the lemma , let @xmath455 in eq .
( [ defd ] ) .
since @xmath452 is hermitian , we can write it in its eigendecomposition @xmath456 with the eigenvalues @xmath457 and the eigenvectors @xmath458 .
each eigenstate @xmath459 can be written in a schmidt decomposition as @xmath460 with @xmath461 , and all vectors normalized .
note that the states @xmath462 and @xmath463 can span a different two - dimensional subspace of @xmath464 for each @xmath72 .
there exists a local filter @xmath465 @xcite from which we can obtain the state @xmath459 from the maximally entangled state @xmath466 : @xmath467 @xmath465 includes : ( 1 ) a unitary transformation from the basis @xmath468 to @xmath469 , where @xmath470 are the schmidt vectors of @xmath471 when it is written in the form @xmath472 ( taking advantage of the degeneracy of the schmidt decomposition of the maximally entangled state ) , and ( 2 ) a diagonal filter which reduces the schmidt coefficients to @xmath473 . thus we may write @xmath452 as @xmath474 we see that since @xmath455 by assumption , we are able to write the action of the map @xmath378 on the input @xmath466 in a ` completely positive form ' with operation elements @xmath475 that depend on @xmath476 .
we observed above that this input determines the action of the map uniquely on the subspace spanned by the vectors @xmath453 and @xmath454 .
therefore the map acts as a completely positive map on any input that has support on a two - dimensional space .
this implies that eq .
( [ anystate ] ) holds for any state @xmath477 .
conversely , if @xmath378 is @xmath362-positive then eq . ( [ 2posmax ] ) holds for any maximally entangled state @xmath466 .
@xmath138 c. h. bennett , p. w. shor , j. a. smolin , and a. v. thapliyal , `` entanglement - assisted classical capacity of noisy quantum channels , '' phys .
lett . , in press ;
eprint quant - ph/9904023 .
a. ekert , phys .
lett . * 68 * , 661 ( 1991 ) .
r. cleve and h. buhrman , phys .
a * 56 * , 1201 ( 1997 ) , eprint quant - ph/9704026 ; r. cleve , h. buhrman , and a. wigderson , in _ proc . of the 30th ann .
acm symp . on the theory of computing
_ ( acm press , 1998 ) , p. 63 , eprint quant - ph/9802040 . c. h. bennett , g. brassard , c. crpeau , r. jozsa , a. peres , and w. k. wootters , phys .
* 70 * , 1895 ( 1993 ) .
d. mayers , phys .
lett . * 78 * , 3414 ( 1997 ) ; h .- k .
lo and h. f. chau , _ ibid . _ * 78 * , 3410 ( 1997 ) ; h .- k . lo and h. f. chau , physica d * 120 * , 177 ( 1998 ) , eprint quant - ph/9711065 .
m. horodecki and p. horodecki , `` reduction criterion of separability and limits for a class of protocols of entanglement distillation '' , phys .
a * 59 * , 4206 ( 1999 ) ; eprint quant - ph/9708015 .
s. l. braunstein , c. m. caves , r. jozsa , n. linden , s. popescu , and r. schack , phys .
* 83 * , 1054 ( 1999 ) , eprint quant - ph/9811018 .
the notation @xmath42 means that alice s hilbert space @xmath464 has dimension @xmath6 , and bob s hilbert space @xmath478 has dimension @xmath349 .
a. peres , phys .
* 77 * , 1413 ( 1996 ) , eprint quant - ph/9604005 .
p. horodecki , phys .
a * 232 * , 333 ( 1997 ) , eprint quant - ph/9703004 .
see m. horodecki , p. horodecki and r. horodecki , phys .
lett . a * 223 * , 1 ( 1996 ) , eprint quant - ph/9605038 .
horodecki , m. horodecki , and r. horodecki , `` bound entanglement can be activated , '' phys
lett , * 82 * , 1056 - 1059 ( 1999 ) , eprint quant - ph/9806058 .
m. horodecki , p. horodecki and r. horodecki , phys .
lett . * 80 * 5239 ( 1998 ) .
monwhea jeng , private communication .
d. p. divincenzo , b. m. terhal , and a. v. thapliyal , `` optimal decompositions of barely separable states , '' j. mod .
optics , in press ; eprint quant - ph/9904005 .
r. b. lockhart , `` optimal ensemble length of mixed separable states , '' eprint quant - ph/9908050 .
some of the literature on positive maps : m. -d .
choi , can .
j. math .
xxiv , no . 3 ( 1972 ) , 520 - 529 ; h. osaka , lin . alg . and its appl .
* 186 * , 45 - 53 ( 1993 ) .
armin uhlmann , private communication , and theorem 2 in j. tomiyama , lin .
alg . and its appl .
* 69 * , 169 ( 1985 ) .
w. dr , j. i. cirac , m. lewenstein , and d. bruss , phys .
a * xx * , xxxxxx ( 2000 ) ; quant - ph/9910022 . | we exhibit a two - parameter family of bipartite mixed states @xmath0 , in a @xmath1 hilbert space , which are negative under partial transposition ( npt ) , but for which we conjecture that no maximally entangled pure states in @xmath2 can be distilled by local quantum operations and classical communication ( lq+cc ) .
evidence for this undistillability is provided by the result that , for certain states in this family , we can not extract entanglement from any arbitrarily large number of copies of @xmath0 using a projection on @xmath2 .
these states are canonical npt states in the sense that any bipartite mixed state in any dimension with npt can be reduced by lq+cc operations to an npt state of the @xmath0 form .
we show that the main question about the distillability of mixed states can be formulated as an open mathematical question about the properties of composed positive linear maps . |
our understanding of atomic theory is largely based on the independent - electron model in which the effect of the inter - electronic repulsion , globally referred to as correlation , is disregarded .
the correlation effects can , however , have major influence on measureable quantities of atomic systems .
the correlation energy , defined by the difference between the exact total energy and hartree - fock energy , is traditionally used as a guide @xcite for the amount of correlation in a given system .
relatively recently , it has been proposed that shannon s information entropies @xcite provide a useful basis for the measure of electron correlation in atomic systems @xcite . for a many - electron atom ,
shannon s position - space information entropy is defined by @xcite @xmath1 where @xmath2 with @xmath3 , the normalized wave function of an @xmath4 electron atom .
the charge density @xmath5 is also normalized to unity .
correspondingly , the momentum- space entropy is written as @xmath6 the momentum - space single - particle density @xmath7 is similar to that given in ( 2 ) but characterized by the normalized momentum - space wave function @xmath8 obtained by taking fourier transform of the position - space wave function .
the two entropies as introduced through ( 1 ) and ( 3 ) allowed bialynicki - birula and mycielski @xcite to introduce a stronger version of the heisenberg uncertainty relation which for any 3-dimensional system reads @xmath9 equation ( 4 ) , often referred to as bbm inequality , clearly indicates the reciprocity between the representation and momentum spaces such that high values of @xmath10 are associated with low values of @xmath11 . for uncorrelated two - electron systems ( separable wave functions )
the entropy sum is independent of the atomic number @xmath0 and has a constant value 6.5665 .
but when the electron - electron interaction is switched on , the entropy sum appears to depend on @xmath0 @xcite .
consequently , the correlated entropy sum @xmath12 has also been used as measure of correlation in atomic systems .
information entropies of two - electron systems can be expressed in closed analytic form for separable approximation of the wave function .
this is , however , not possible for explicitly @xmath13 - ( inter - electronic separation ) dependent wave functions which can account for correlation in the system.the object of the present work is to derive an exact analytic model to quantify correlation in two - electron systems described by wave functions depending explicitly on @xmath13 . in particular
, we shall use our model to study the effect of inter - electronic repulsion on shannon s information entropies with a view to look for a new measure of correlation .
the system of our interest consists of the first five members in the helium iso - electronic sequence .
there exist elaborate studies on the problem of determining accurate wave functions for the helium atom @xcite .
the constructed wave functions often involve very large number of variational parameters .
consequently , it is rather difficult to extend such approaches to deal with other ions in the helium iso - electronic sequence .
moreover , in applicative context , use of these wave functions leads to very large number of computations . in view of this , occasionally , there have been attempts to construct accurate correlated two - electron wave functions involving fewer numbers of parameters .
one such attempt is due to bhattacharyya et al .
an important virtue of their work is that the proposed three - parameter correlated wave function could be used to construct exact analytic expression for the total energy of two - electron systems .
but it was noted that the wave function of ref .
7 satisfies cusp conditions @xcite only approximately such that the total energy computed for the helium atom is higher than the experimental value . as a result
there have been many attempts to improve on the quality of such few - parameter wave functions .
a significant work along this line of investigation is due to le sech @xcite who constructed a two - parameter wave function which not only satisfies the correlation cusp condition to a high degree of accuracy but also gives much improved values for the energies of two - electron atomic / ionic systems from @xmath14 to @xmath15 .
recently , chauhan and harbola @xcite reexamined the optimization procedure for the parameters in the wave function of le sech in order to make it more accurate than that used in ref.9 for the computation of ground - state energies of two - electron atoms and ions .
the wave function of our interest can be written in the form @xmath16 with @xmath0 , the atomic number of the system and @xmath17 , the normalization constant of the wave function . here
@xmath18 and @xmath19 are variational parameters . for @xmath20 ( 5 )
gives the wave function of le sech @xcite while for @xmath21 and @xmath22 we get the well known separable two - electron wave function @xcite .
we shall construct closed form expressions for both single - particle coordinate- and momentum - space charge densities and then employ them to examine effect of correlation on individual entropies and entropy sum @xmath12 by using the parameters of chauhan and harbola @xcite .
we begin sec .
ii by noting that instead of using ( 2 ) and a similar result written in the momentum space , it is possible to obtain expressions for single - particle charge densities by following an alternative approach .
the second viewpoint is particularly well suited to derive results for position- and momentum - space charge densities @xmath5 and @xmath7 in the presence of correlation .
we demonstrate that , due to the effect of inter - electronic repulsion , the bare or uncorrelated radial charge distribution is pushed apart . on the other hand ,
the bare momentum distribution is squeezed by the same effect . in sec .
iii we present results for the position- and momentum - space information entropies computed on the basis of our expressions for @xmath5 and @xmath7 .
we verify that the uncorrelated entropy sum is a constant which does not depend on the atomic number @xcite .
as opposed to this , in the presence of correlation the entropy sum becomes @xmath0 dependent .
however , correlated entropy sum tends to the un - correlated @xmath0-independent constant value as we go along the iso - electronic sequence .
the position - space entropy @xmath10 takes up negative values for high-@xmath0 atoms but values of momentum - space entropy @xmath11 are always positive .
we provide a geometrical realization for this fact and thus try to visualize how the interplay between nuclear coulomb interaction and inter - electronic repulsion determines the variation of @xmath10 or @xmath11 as a function of @xmath0 .
we observe that individual entropies are more strongly affected by the inter - electronic repulsion than the corresponding entropy sum such that the shannon s information entropy is a very suitable candidate for the measure of correlation . finally , we make some concluding remarks in sec .
we shall first work with an uncorrelated or separable wave function for the two - electron system and show that , rather than using ( 2 ) or its momentum - space analog , it is possible to follow a different route to construct expressions for charge densities .
the well known two - electron separable wave function is given by @xcite @xmath23 with @xmath24 an appropriate normalization constant . from ( 2 ) and ( 6 ) it is straightforward to show that @xmath25 represents the normalized single - particle charge density of the system .
the result in ( 7 ) could also be derived by first constructing a single particle wave function @xmath26 by integrating ( 6 ) over @xmath27 and subsequently normalizing the wave function in ( 8) to unity .
it is straightforward to verify that the alternative approach presented above is equally applicable for the momentum - space ( @xmath28-space ) separable wave function @xmath29 obtained by taking the fourier transform of ( 6 ) . here
@xmath30 is the normalization constant of the momentum - space wave function .
this implies that the momentum - space charge density @xmath7 computed from the wave function @xmath31 is exactly equal to that found from the momentum - space analog of ( 2 ) .
we shall now follow the second viewpoint to derive expressions for charge densities of correlated two - electron systems described by non - separable wave functions . to that end
we introduce , for brevity @xmath32 , and decompose ( 5 ) as a sum of two parts @xmath33 such that @xmath34\ ] ] and @xmath35r_{12}e^{-br_{12}}.\ ] ] equation ( 11 ) can easily be integrated over the variable @xmath27 to get @xmath36.\ ] ] it is , however , nontrivial to integrate ( 12 ) because @xmath37 occurs here as an entangled variable .
this poses an awkward analytical constraint to derive an expression for the single particle charge density by using non - separable wave functions .
fortunately , we can make use of the identity @xcite @xmath38 to circumvent the difficulty . for example
, from ( 14 ) we can deduce that @xmath39 which when substituted in ( 12 ) provides a separable representation of @xmath40 such that @xmath41 with @xmath42 and @xmath43 equations ( 17 ) and ( 18 ) can be integrated over the variable @xmath37 by making use of @xcite @xmath44 this gives @xmath45 with @xmath46 and @xmath47 the angular integrals in ( 21 ) and ( 22 ) can be evaluated with the help of @xmath48 and , finally integration over @xmath49 yields @xmath50 and @xmath51 as with ( 9 ) , the results in ( 13 ) and ( 20 ) together with ( 24 ) and ( 25 ) give a single - electron wave function corresponding to the non - separable or entangled state ( 5 ) .
understandably , the present single - electron wave function obtained by eliminating @xmath27 from the two - particle one will involve the effect of correlation
. the uncorrelated wave function can be obtained from ( 13 ) and ( 20 ) using @xmath21 and @xmath22 .
the integral in ( 19 ) in conjunction with @xmath52 and @xmath53 can now be used to obtain the momentum - space wave function from our constructed coordinate - space result .
both coordinate- and momentum - space wave functions involve normalization constants .
if we fix their values by normalizing the corresponding wave functions to unity , we can write expressions for coordinate- and momentum - space charge densities . in this context
it is important to note that in our approach we could study the momentum - space properties of two - electron systems using tabulated integrals only . in the past
such studies were found to call for the use of monte carlo algorithms to calculate the integrals involved @xcite .
it is also straightforward to compute numbers for @xmath10 and @xmath11 with the help of our expressions for charge densities . in this context
we observe that , in contrast to the case of correlated two - electron systems , it is relatively uncomplicated to construct closed form analytic expressions for position- and momentum - space charge densities as well as corresponding information entropies for the one - electron atom @xcite .
the polynomials occurring in the wave functions of hydrogenic excited states , however , tend to present difficulties to compute results for associated information entropies .
one of us @xcite demonstrated that the problem can be circumvented by the simultaneous use of series and product representations of the polynomials @xcite .
we shall first examine the effect of inter - electronic repulsion on the charge - densities by using the parameters of the wave function ( 5 ) as given in ref .
10 . in table 1
we present numbers for correlated and uncorrelated charge distributions at @xmath54 ( coordinate space ) and @xmath55 ( momentum space ) for the two - electron systems from @xmath14 to @xmath56 .
for ready reference , we also include in this table the parameters of the wave function .
understandably , the uncorrelated charge distributions are obtained from the corresponding correlated distributions using @xmath21 and @xmath22 .
.parameters of the wave function ( 5 ) and values of the coordinate- and momentum - space charge densities at @xmath54 and @xmath55 . [ cols="<,<,<,<,<",options="header " , ] looking closely into these numbers we see that the position - space entropies @xmath57 and @xmath58 are decreasing function of @xmath0 while @xmath59 and @xmath60 are increasing function of @xmath0 .
recently , while examining the role of statistical correlation in the n - particle moshinsky model , peng and ho @xcite observed a similar behavior of shannon information entropy in different phase spaces . in two interesting publications
koscik and okopinska @xcite studied the behavior of von neumann and linear entropies in the moshinsky model and in helium - like atoms .
it is rather interesting to note that the response of these entropies to inter - electronic repulsion appears to agree with that exhibited by the position - space shannon entropies in table 2 . as for the effect of inter - electronic repulsion on the individual entropy we note
that correlation plays an opposite role in modifying the values of @xmath57 and @xmath59 .
for example , effect of correlation always increases the value of position - space entropies .
on the other hand , values of the momentum - space entropies are reduced by inter - electronic repulsion . from the results in column 4
, we see that the entropy sum @xmath61 for the separable wave function is independent of the atomic number @xmath0 .
this was shown analytically in @xcite . from the numbers in column 7 we see that the effect of inter - electronic repulsion increases the entropy sum by about 3 percent for @xmath14 while for @xmath56 a similar increase in the entropy sum is 0.25 percent only .
in fact , the correlation contribution to the entropy sum decreases monotonically as we go along the iso - electronic sequence such that for high @xmath0 values the correlated entropy sum tends to coalesce with the uncorrelated entropy sum .
it is of interest to note that , rather than the entropy sum , the individual entropies @xmath57 and @xmath59 are more significantly affected by inter - electronic repulsion .
for example , @xmath58 of @xmath14 is greater than the corresponding result for @xmath57 by about 11.9 percent .
for the momentum - space entropy of @xmath14 we note that@xmath62 by approximately 12.2 percent . in this context
it may be interesting to note that for helium the correlation correction in @xmath58 is about ten times of what one needs to obtain the experimental binding energy of @xmath63 from the corresponding hartree - fock energy .
thus we would venture to suggest that , rather than the entropy sum or binding energy of two - electron systems , the individual shannon s entropies should be regarded as better candidates for the measure of correlation .
we shall now make use of the plots of entropy densities @xmath64 and @xmath65 as a function of @xmath66 and @xmath28 respectively to provide a physical feeling for why the values of @xmath57 and @xmath58 become negative for high value of @xmath0 while the numbers for @xmath59 and @xmath60 always remain positive . to that end
we first remember that contribution to entropies as found by integrating ( 28 ) and ( 29 ) comes from both core and valence regions in the system . in the core region , the effect of nuclear charge dominates and tends to localize the charge distribution . on the other hand , the inter - electronic repulsion in the valence region causes the density distribution to be delocalized .
the interplay between localization and delocalization is responsible for change in sign in the values of @xmath10 . and
@xmath67 of @xmath14 ion as a function of @xmath66.,width=245 ] we display in fig .
3 the position - space uncorrelated and correlated entropy densities @xmath68 and @xmath67 for @xmath14 as a function of @xmath66 . we shall follow this convention for other atoms also .
the solid and dotted curves denote the variation of @xmath68 and @xmath67 respectively . from the curves in this figure it is evident that the results for entropy densities are positive for all values of @xmath66 .
physically , this implies that contribution to entropies comes predominantly from inter - electronic repulsion and the core contribution is really insignificant .
this is understandable because @xmath14 is a highly delocalized system .
from the areas of the solid and dotted curves below the axis it can be shown that @xmath69 .
this observation is consistent with the results for @xmath14 in table 2 .
the deviation between solid and dotted curves increases at large radial distances where the effect of correlation becomes relatively more important .
the curves in this figure are not discernible from those reported by sagar and guevara @xcite . and @xmath67 of @xmath63 ion as a function of @xmath66.,width=245 ] in fig.4 we present similar results for entropy densities for @xmath63 .
from this figure we see that both solid and dotted curves show small positive peaks near @xmath54 , the peak of the solid curve being little more pronounced .
the positive peaks , however small , arise due to localization of the charge density induced by the electron - nucleus interaction . here
the uncorrelated and correlated entropies result from the cancellation between areas of the curves above and below the @xmath66 - axis .
an important point to note here is that the deviation between the solid and dotted curves has reduced considerably when compared with the deviation between the corresponding curves of @xmath14 . as a result @xmath70 for all @xmath66
this is physically realizable because the effect of correlation diminishes as we move to high @xmath0 atoms .
the positive peaks in the curves arising due to interaction of the electrons with the nuclear charge become more and more pronounced for high @xmath0 atoms . to visualize this
we portray in fig.5 the entropy densities for @xmath56 as a function of @xmath66 . here
we see that the positive peaks are really very much pronounced . and
@xmath67 of @xmath56 ion as a function of @xmath66.,width=245 ] the entropies @xmath71 and @xmath72 for the triply ionized boron are obtained by the algebraic sum of areas of the solid and dotted curves above and below the @xmath66- axis . since the areas of the curves above the @xmath66 axis are bigger than those below the @xmath66 axis , the values of both uncorrelated and correlated entropies are negative .
as expected due to the effect of correlation the number for @xmath58 is greater than that for @xmath57 . here
deviation between the solid and dotted curves is appreciable near @xmath73 only since the electron - nuclear interaction is quite large for @xmath74 . on the other hand , because of weaker correlation effect , the solid and dotted curves exhibit very little deviation for large @xmath66 values . in the following we present a similar analysis for numbers of the momentum - space entropies using the plots of appropriate entropy density functions .
in fig.6 we show @xmath75 and @xmath76 for @xmath14 as a function of @xmath28 . as before solid and
the dotted curves give the variation of uncorrelated and correlated results . and
@xmath76 for @xmath14 versus @xmath28.,width=245 ] in close analogy with the solid curve in fig.3 , the curve for the uncorrelated momentum entropy density is negative for all values of @xmath28 .
but for small values of @xmath28 the dotted curve exhibits a small positive peak then takes up negative values .
the negative part of the dotted curve appears to be slightly squeezed compared to the solid curve . clearly ,
appearance of the positive peak and observed squeezing account for why the effects of electron - nucleus interaction and of correlation reduce the value of the bare momentum - space entropy .
since the effect of electron - electron interaction gradually diminishes and electron - nucleus interaction increases as we go to high @xmath0 atoms , the height of the positive peak in the dotted curve is likely to be modified at large @xmath0 values . to see this we display in
fig.7 the momentum - space entropy density of @xmath63 for which the electron - nucleus interaction is little stronger than the corresponding interaction in @xmath14 . and @xmath76 for @xmath63 versus @xmath28.,width=245 ] and @xmath76 for @xmath15 versus @xmath28.,width=245 ] from this figure
we see that the positive peak in the dotted curve has disappeared due to increase in the strength of the electron - nucleus interaction . in the context of position - space entropy densities we found that the electron - nucleus interaction increases the height of peaks near @xmath54 . here
we observe the opposite .
in contrast to the plot in fig .
6 , the dotted curve here does not have any peak .
this is , however , quite expected since we are now in a reciprocal space .
the nature of the solid and dotted curves as appears in fig.7 does not change for higher atoms in the helium iso - electronic sequence .
this can be seen clearly from the curves in fig.8 giving the plots of momentum - space entropy densities for @xmath15 .
however , here the area of the solid or dotted curve below the axis is much bigger than those in fig.7 such that the results for both uncorrelated and correlated entropies of @xmath15 are greater than the corresponding values for @xmath63 or @xmath14 . here
the deviation between the solid and dotted curves is not appreciable .
consequently , the correlated entropy value differs from the uncorrelated result in table 3 by about 2 percent only .
it is widely believed that there are distinct advantages to viewing problems of physics within the framework of analytical models , since many physical effects are then readily expressed and evaluated .
but in atomic and molecular physics , hardly there are problems that can be solved by using analytical tools .
for example , one can not use realistic atomic wave functions to express single - particle charge densities in closed analytic form for even two - electron systems . in the present paper we work with a highly accurate hylleraas - type two - electron wave function and
derive an analytical model to study the effect of correlation on the single - particle position- and momentum - space charge densities , as well as corresponding shannon information entropies . with special attention to the first five members of the helium iso - electronic sequence we verify that inter - electronic repulsion flattens the position - space charge density and squeezes the momentum - space charge density .
as expected , the effect of correlation is more pronounced for low-@xmath0 ions .
the expressions for single - particle charge densities are used to compute numbers for shannon s position- and momentum - space information entropies @xmath10 and @xmath11 respectively . both uncorrelated and correlated
values of @xmath10 are decreasing function of @xmath0 while those of @xmath11 have been found to increase with the atomic number .
but the relative contribution to @xmath10 due to inter - electronic repulsion increases as we go to high @xmath0 atoms .
but for @xmath11 we observe the opposite .
recently , lin and ho @xcite made use of a hylleraas - type wave function involving more than 400 parameters to obtain numbers for shannon position - space information entropies by using a purely numerical routine .
the results obtained by them for @xmath14 , @xmath63 and @xmath77 , although somewhat augmented to our results , support our viewpoint for the effect of correlation on @xmath57 .
but the work in ref.22 could not be extended to deal with the corresponding momentum - space problem .
we attribute the observed variation of @xmath10 or @xmath11 with atomic number to the interplay between interactions in the core and valence regions of the atom and thus provide a geometrical realization for the variation .
more significantly , it is shown that correlation plays a more dominant role in modifying the bare information entropies than it does in correcting the entropy sum as well as hartree - fock binding energies of atomic systems . in view of the observed overall exaggerated response of shannon information entropies to inter - electronic repulsion
we suggest that either @xmath10 or @xmath11 could be used as a correlation measure . in the literature
there are other useful suggestions for electron correlation tools .
for example , romera and dehesa @xcite introduced a measure of correlation in two - electron systems by analyzing the product of shannon entropy power and the fisher information of the electron density . almost simultaneously with this work , huang and kais @xcite proposed entanglement as an alternative measure of electron correlation .
since entanglement is one of the most striking phenomena of quantum mechanics and , at the same time , is directly observable , the work in ref .
24 deserve some special attention .
a few years ago , some serious attempts were made to construct accurate but simple correlated wave functions for a large number of two - electron systems @xcite .
there also exist simple correlated wave functions for the k- shell electrons of many neutral atoms @xcite .
the wave functions in these works differ among themselves only in their dependence on the inter - electronic separation . the analytical model developed in this paper
can easily be extended to work with each of these two - electron wave functions .
it will , therefore , be interesting to apply our analytical model to systems considered in refs .
13 and 25 to 27 in order to provide further evidences for exaggerated response of shannon s information entropies to inter - electronic repulsion .
99 p. o. lwdin , quantum theory of many - particle systems i. physical interpretation by means of density matrices , natural spin - orbitals , and convergence problems in the method of configuration interaction , phys .
rev . * 97 * , ( 1955 ) 1447 - 1589 ; quantum theory of many - particle systems ii .
study of the ordinary hartree - fock approximation , phys . rev . * 97 * , ( 1955 ) 1490 - 1508 ; quantum theory of many - particle systems iii .
extension of the hartree - fock scheme to include degenerate systems and correlation effects , phys .
97 * , ( 1955 ) 1509 - 1520 . c. e. shannon , a mathematical theory of communication , bell syst .
j. * 27 * , ( 1948 ) 623 - 656 .
p. ziesche , v. h. smith jr . , m. ho , s. p. rudin , p. gersdorf and m. taut , the he isoelectronic series and the hook s law model : correlation measures and modifications of collinsc onjecture , j. chem
. phys . * 110 * , ( 1999 ) 6135 - 6142
. i. bialynicki - birula and j. myceilski , uncertainty relations for information entropy in wave mechanics , commun . math .
phys . * 44 * , ( 1975 ) 129 - 132 .
n. l. guevara , r. p. sagar and r. o. esquivel , local correlation measures in atomic systems , j. chem .
* 122 * , ( 2005 ) 084101 8pp .
e. a. hylleraas , uber den grundterm der zweielektronenprobleme von @xmath14 , @xmath63 , @xmath77 , @xmath78 @xmath79 , z. phys .
* 65 * , ( 1930 ) 209 - 225 ; s. chandrasekhar , some remarks on the negative hydrogen ion and its absorption coefficient , astrophys .
j. * 100 * , ( 1944 ) 176 - 180 ; t. kinoshita , ground state of the helium atom ii , phys . rev .
* 115 * , ( 1959 ) 366 - 374 .
s. bhattacharyya , a. bhattacharya , b. talukdar and n. c. deb , analytical approach to the helium - atom ground state using correlated wavefunctions , j. phys .
b : at . mol .
* 29 * , ( 1996 ) l147 - l150 .
c. c. j. roothaan and a. w. weiss , correlated orbitals for the ground state of helium like systems , rev .
phys . * 32 * , ( 1960 ) 194 - 205 . c. le sech , accurate analytic wavefunctions for two - electron atoms , j. phys .
b : at . mol .
. phys . * 30 * , ( 1997 ) l47 - l50 .
r. s. chauhan and m. k. harbola , highly accurate wavefunction for two - electron systems using two parameters , arxiv : 1506.00912v2 [ physics .
atom - ph ] 3 jun 2015 .
l. i. schiff , _ quantum mechanics _ tata mcgraw hill , new delhi , india , 2010
. g. b. arfken , _ mathematical methods for physicists _ , second ed . , academic press , new york , 1970 .
a. sarsa , f. j. galvez and e. buendia , correlated two - electron momentum properties of helium to neon atoms , j. chem .
phys . * 110 * , ( 1999 ) 5721 - 5727 .
j. s. dehesa , w.v .
asssche and r. f.yanes , information entropy of classical orthogonal polynomials and their application to the harmonic oscillator and coulomb potential , methods and application of analysis * 4 * , ( 1997 ) 91 - 110 .
a. bhattacharya , b. talukdar , u. roy and angshula ghosh , quantal information entropies for atoms , int . j. theor . phys . * 37 * , ( 1998 ) 1667 - 1680 .
m. abramowitz and i. a. stegun , _ handbook of mathematical functions _
, dover , new york , 1970 .
t. koga , y. kasai and a. j. thakkar , accurate algebraic densities and intracules for helium - like ions , int .
. chem . * 46 * , ( 1993 ) 689 - 699 .
a. j. thakkar and s. p.
mc carthy , electron and electron - pair number and momentum densities , adv .
quant . chem .
* 67 * , ( 2013 ) 19 - 54 .
h. t. peng and y. k. ho , statistical correlation of the n - particle moshinsky model , entropy * 17 * , ( 2015 ) 1882 - 1895 .
p. koscik and a. okopinska , correlation effects in moshinsky model , few - body syst . *
54 * , ( 2013 ) 1637 - 1640 ; entanglement entropies in the ground states of helium - like atoms , few - body syst . *
55 * , ( 2014 ) 1151 - 1157 .
r. p. sagar and n. i. guevara , mutual information and electron correlation in momentum space , arxiv : quant- ph//0602046v1 3 feb 2006 . c. h. lin and y. k. ho , shannon information entropy in position space for two - electron atomic systems , chem .
. lett . * 633 * , ( 2015 ) 261 - 264 .
e. romera and j. s. dehesa , the fisher - shannon information plane , an electron correlation tool , j. chem .
* 120 * ( 2004 ) 8906 - 8912 . z. huang and s. kais , entanglement as a measure of electron - electron correlation in quantum chemistry calculations , chem .
phys . letts . *
412 * , ( 2005 ) 1 - 4 .
k. v. rodriguez , g. gasanco and d. m. mitnik , accurate and simple wavefunctions for the helium isoelectronic sequence with correct cusp - conditions , j. phys .
b : at . mol .
* 40 * , ( 2007 ) 3923 - 3939 .
f. otranto , g. gasaneo and c. r. garibotti , nucl .
energy and cusp conditions study for the he isoelecctronic sequence , nucl .
instr . and methods in phys .
b * 217 * , ( 2004 ) 12 - 17 .
d. m. mitnik and j. e. miraglia , simple correlated wavefunctions for the k - shell electrons of neutral atoms , j. phys .
b ; at . mol . opt
. phys . * 38 * , ( 2005 ) 3325 - 3338 ; k. v. rodriguez , g. gasaneo , d. m. mitnik and j. e. miraglia , hylleraas - like functions with correct cusp conditions : k - shell electrons for neutral atoms , j. elect
. spec . * 151 * , ( 2007 ) 204 - 206 . | we make use of a hylleraas - type wave function to derive an exact analytical model to quantify correlation in two - electron atomic / ionic systems and subsequently employ it to examine the role of inter - electronic repulsion in affecting ( i ) the bare ( uncorrelated ) single - particle position- and momentum - space charge distributions and ( ii ) corresponding shannon s information entropies .
the results presented for the first five members in the helium iso - electronic sequence , on the one hand , correctly demonstrate the effect of correlation on bare charge distributions and , on the other hand , lead us to some important results for the correlated and uncorrelated values of the entropies .
these include the limiting behavior of the correlated entropy sum ( sum of position- and momentum - space entropies ) and geometrical realization for the variation of information entropies as a function of @xmath0 .
we suggest that , rather than the entropy sum , individual entropies should be regarded as better candidates for the measure of correlation . |
it is well known that precision and accuracy of astronomical observations , both optical and radio , made through the earth s atmosphere depend on the elevation at which the object is observed .
these errors grow with decreasing of the elevation due to larger air mass and difficulties in modelling of refraction effects at low elevation . from this point of view observations
should be made in the near - zenith zone when possible .
on the other hand , inclusion in processing of observations made at low elevations is important when definite groups of highly correlated parameters , for instance station coordinates and zenith troposphere delays , are estimated simultaneously .
in such a case using observations made at in a widest range of elevation allows us to mitigate the correlations between unknowns and improve the solution . to meet these mutually exclusive requirements , proper elevation - dependent weighting ( edw ) of observations is used . in a special case of step - like weighting function , i.e. rejection of the observations made at the elevation less than the given limit
, such a limit usually is called cut - off elevation angle ( cea ) .
it was shown in many studies that elevation - dependent weighting may have a significant impact on the results of processing of the space geodesy observations .
in particular , several studies of this effect was made by the goddard and vienna vlbi analysis groups in the framework of the ivs vlbi2010 committee activity .
they investigated an influence of cea and edw on geodetic results such as earth orientation parameters ( eop ) , baseline length repeatability , troposphere parameters , and station heights .
those results were based on simulation .
gipson in @xcite used another approach to edw .
he applied elevation dependent additive noise to the measurement error instead of using a weighting factor as it is usually being made .
he tested his method with the actual cont05 observations .
results of both mentioned and other results are sometimes contradictory .
this gave an impulse to the present work where some results are presented of investigation of the impact of the cea and edw on the baseline length repeatability and eop estimates .
for this test , cont05a observations were processed making use of occam software with different edw functions including cea , keeping all other options as follows : * kalman filter mode ( kf ) , * random walk model for clocks , psd=1.5 ps@xmath0/s , * random walk model for ztd , psd=0.25 ps@xmath0/s , * one ns and ew troposphere gradient estimate for the session . for the continuous edw mode ( continuous weighting function ) , the measurement error is multiplied by a factor @xmath1 where @xmath2 and @xmath3 are edw parameters , @xmath4 is the source elevation .
such a weighting function provides a smooth stepless change in weight for any @xmath2 .
one can see that a case of @xmath5 gives merely @xmath6 , which is close to actual wet mapping function used in the last works by macmillan and gipson ( private communications ) .
figure [ fig : mf ] shows actual hydrostatic and wet mapping functions along with approximation @xmath7 for a typical cont05 session .
one can see that these functions are close enough , and either of them can be used for the edw without significant impact on the result .
= 0.48 = 0.48 edw mode with @xmath8 was implemented in the occam / gross software @xcite for routine data processing .
it will be referred hereafter as `` normal mode '' . in a case of cea
we have @xmath9 the latter line corresponds to the kf realization used in occam . for vlbi delay , measurement error coming from correlator is multiplied by two @xmath10 values computed for both the stations . in our test
, @xmath11 were used for cea test , and @xmath12 were used for continuous edw mode .
test results obtained with different cea are shown in figure [ fig : cea_baselines ] .
the case of @xmath13 includes all the observations without weighting , since no cont05 observations were made at the elevation less than 4@xmath14 .
= 0.48 + = 0.48 table [ tab : edw_baseline ] shows edw test results .
different edw modes are denoted as w_e_p , where e and p are @xmath2 and @xmath3 in eq.([eq : edw ] ) .
test results are given for quadratic approximation in percent with respect to the case of cea with @xmath13 .
one can see that several edw modes show about the same improvement in the baseline length repeatability .
.comparison of the baseline repeatability obtained with different edw models .
see explanation in text . [
cols="^,^,^,^,^ " , ]
the preliminary conclusions from this test are the following . *
the baseline length repeatability steadily grows with the cea increasing , remaining practically the same in the cut - off angle range from 3@xmath14 ( i.e. no cut - off for the cont05 ) to 9@xmath14 . *
the best result is obtained when the edw elevation - depending weighting is applied to the low - elevation observations .
however , the test results are not always unambiguous .
further adjustment of the weighting method may be fruitful . *
the xp , yp and ut1 uncertainties grow with the increasing cut - off angle after about 10@xmath14 . most probably , this reflects the fact that only about 6% of the total number of cont05 observations were made at the elevations below 10@xmath14 .
the xc and yc uncertainties and scatter depend on the cea much less . *
xp bias w.r.t .
igs slightly depends on the cea , except the maximum tested cea values , evidently unrealistic .
in contrast , yp bias substantially changes with increasing cea .
most probably , this can be explained by the cont05 network orientation , for which the longitude of the central meridian @xmath15 just corresponds to the y direction of the terrestrial coordinate system .
* some statistics such as the uncertainty and the scatter of the xc and yc , as well as the wrms of xp and yp w.r.t .
igs have the minimum at the cea around @xmath16 , which is interesting and deserves a supplement investigation . *
as one can expect , the correlations between eop comprising xp and yp grow with increasing cea , but remain small due to good cont05 network geometry .
the same can be expected for the ivs2010 network .
the correlation between xc and yc remain practically the same for all tested cea , except the maximum tested cea value , evidently unrealistic .
finally , we can conclude that inclusion of the low - elevation observations , properly weighted , improves the baseline length repeatability and eop results . on the contrary ,
filtering the observations using the cut - off elevation method may lead to degradation of geodetic results .
however , this should be mentioned that the conclusions drawn from the result obtained in this paper has been proven with standard geodetic vlbi observations , where rather few observations were made at low elevations , as mentioned above .
perhaps , special r&d sessions with more uniform distribution of observations over the sky , including observations at very low elevations , may be useful for more detailed study on the impact and optimal processing of the low - elevation observations on geodetic parameters obtained from vlbi observations .
it ought be mentioned that all the edw modes considered in this paper in fact modify only diagonal elements of the corresponding covariance matrix . according to gipson
s work @xcite best result can be achieved in a case of account also for correlations between observations .
it seems to be interesting to investigate how this approach will work in kf estimator .
gipson , j. incorporating correlated station dependent noise improves vlbi estimates . in : j. boehm , a. pany , h. schuh ( eds . ) , proc .
18th european vlbi for geodesy and astrometry working meeting , vienna , austria , 12 - 13 apr 2007 , geowissenschaftliche mitteilungen , heft nr .
79 , schriftenreihe der studienrichtung vermessung und geoinformation , technische universitaet wien , 2007 , 129134 . | in this paper , results are presented on studies which have been performed to investigate the impact of the cut - off elevation angle ( cea ) and elevation - dependent weighting ( edw ) on the eop estimates and baseline length repeatability . for this test
, cont05 observations were processed with different cea and edw , keeping all other options the same as used during the routine processing .
uncertainties and biases , as well as correlations between estimated parameters have been investigated .
it has been shown that small cea , up to about 810 degrees does not have large impact on the results , and applying edw allows us to get better result ( smaller errors ) . however
, this result has been proven with standard geodetic vlbi observations , where rather few observations were made at low elevations .
perhaps , special r&d sessions with more uniform distribution of observations over elevation may be useful for more detailed study on the subject . ' '' '' width 0.4 0.2ex 5th ivs general meeting , st .
petersburg , russia , 36 march 2007 |
the skyrme model @xcite is an example of what qcd might look like in the classical or large number of colors ( @xmath0 ) limit @xcite .
the dynamics of the @xmath1 skyrme model is carried purely by a classical pion field .
hadrons appear as topological solitons in this non - linear meson field theory .
these are the appropriate degrees of freedom for the non - perturbative , long wavelength limit of qcd , and hence for low energy baryon and pion physics .
the skyrme model has been applied to the nucleon static properties @xcite and the nucleon - nucleon interactions @xcite with reasonable success . in the last few years
, nucleon annihilation has been investigated from the skyrme point of view .
sommermann et al .
@xcite studied the dynamics of ungroomed skyrmion - antiskyrmion ( @xmath2 ) collisions .
they found that annihilation proceeds quickly with the creation of a coherent pion pulse .
this was confirmed by shao , walet and amado @xcite .
the notion that annihilation leads to an intense coherent pion field burst gives reason for considering annihilation within the classical skyrme approach .
the idea , that the annihilation products , pion and other mesons , come from a coherent wave of meson fields arising from soliton - antisoliton dynamics , turns out to be very fruitful @xcite .
experimental data such as annihilation branching ratios among meson types and pion charge types from low energy annihilation are well explained with minimal parameters@xcite .
furthermore this picture provides a unified view of annihilation in which all the channels come from a single process .
previous studies of annihilation in the skyrme context , with the exception of @xcite , have concentrated on the final state mesons . a full account of the process requires a description of the initial state nucleon - antinucleon interaction and of the dynamics leading up to annihilation as well .
in this paper , as a first step in that directions , we extend the application of the skyrme model to the interaction of @xmath3 in the product ansatz .
( we note that the energy of @xmath2 in the product ansatz was studied for two configurations by musakhanov and musatov @xcite . )
phenomenologically , the @xmath3 potential is not as well established as the @xmath4 potential . at distances
less than one fermi , the interaction is dominated by annihilation . however , at larger distances , a meaningful potential can be defined and studied either by @xmath5-transformation on the @xmath4 meson exchange potential or phenomenologically . here
we will compare our skyrme model results to this phenomenology .
we will see that at large distance , where the product ansatz makes the best sense , the potentials we find agree qualitatively and , in most cases , quantitatively with phenomenological interactions . at intermediate and short distances , we do less well , but at these distances the product ansatz is not valid .
however it is still suggestive . to obtain the interaction at intermediate distances , we would need to study the full skyrme dynamics at these distances .
for the static @xmath2 this is somewhat more complex than in the corresponding @xmath6 case studied by walhout and wambach , , but is possible and we plan to return to it .
the full , time dependent , dynamical @xmath2 problem is far more difficult than the @xmath6 case and is plagued for @xmath2 by numerical instabilities @xcite .
for all these reasons , and because this paper is a first step , we begin by exploring the interaction in the product ansatz . in order to carry out our study
it is necessary to include @xmath7 and @xmath8 mixing in the potential as was first suggested in the @xmath4 case @xcite . in the context of the product ansatz
, we find that the ungroomed @xmath2 channel studied in @xcite is the most attractive channel and that it leads to rapid annihilation . for non - zero grooming
, we find that the @xmath2 interaction can be repulsive .
therefore it seems likely that the dynamics in groomed channels could be very different from that exhibited in @xcite
. since the physical nucleon is represented by an average over differently groomed skyrmions , annihilation in the nucleon - antinucleon system may proceed more slowly than that of ungroomed @xmath2 . in sec . 2 we study the interaction energy of @xmath2 as a function of separation and relative grooming in the product ansatz .
we start by studying the very simplified case of skyrmion and groomed antiskyrmion on top of each other .
this is a physically artificial case , but it permits analytic evaluation and teaches us something , albeit qualitative , about the dependence of the @xmath2 interaction on grooming . we find that for zero separation and zero grooming , the @xmath2 system has zero total energy , as we expect .
this is the case of complete annihilation .
however at relative grooming angle of @xmath9 , and still zero separation , the product ansatz gives a total energy for the @xmath2 system of four times the single skyrmion energy , corresponding to a very repulsive interaction .
this clearly indicates that the @xmath2 interaction is a strong function of relative grooming .
next we study , always in the product ansatz , the interaction energy of @xmath2 at non - zero separation as a function of grooming .
we project to the nucleon space by the algebraic methods of @xcite which also include finite @xmath0 corrections . in sec .
3 we consider the effects of rotational excitations by including intermediate states with @xmath7 and @xmath8 .
we first evaluate the corrections to the @xmath3 potential in perturbation theory and then study the effect fully by diagonalizing in the space spanned by @xmath10 , @xmath7 and corresponding antiparticles .
our results are presented in sec .
we calculate the energy of the skyrmion and antiskyrmion system using the skyrme lagrangian .
the density of this lagrangian is given by @xmath13 where @xmath14 is a unitary @xmath1 valued field and @xmath15.\ ] ] the first term in the lagrangian comes from the non - linear @xmath16-model and the second is the skyrme term .
the third term is a pion mass term and we take @xmath17 mev .
we take the parameters in the lagrangian to have the values @xmath18 mev and @xmath19 .
these values guarantee that the long distance tail of the nucleon - antinucleon interaction will agree with phenomenology , by virtue of the goldberger - treiman relation .
we begin by studying the energy in the product ansatz for the case of zero separation .
we include the first two terms in the lagrangian . the mass term is neglected since it does not lead to additional understanding of this simple configuration .
let @xmath20 be the ungroomed skyrme @xmath1 field .
the ungroomed @xmath12 would be @xmath21 .
the rotation or grooming matrix @xmath22 on @xmath12 is @xmath23 this corresponds to a grooming rotation through angle @xmath24 about the @xmath25 axis .
the product ansatz with this relative grooming is @xmath26 note that the energy is a function of only the relative grooming and should be zero with no grooming ( since for @xmath27 , @xmath28 ) .
the energy density is @xmath29 ^ 2\ ] ] in skyrme units ( energy in @xmath30 and length in @xmath31 ) .
the chiral ( left handed ) derivative is @xmath32 suppose for the skyrme chiral angle @xmath33 , the mass contribution to the @xmath34 skyrmion from the non - linear @xmath35-model term is @xmath36 and from the skyrme term @xmath37 .
after some algebra , we arrive at the following result for the total energy of the @xmath2 product ansatz at relative grooming angle @xmath24 and zero separation .
@xmath38 for the profile @xmath33 which minimizes the @xmath34 skyrme mass , we have @xmath39 from scaling arguments . here
@xmath40 is the mass of the skyrmion .
we then have @xmath41,\ ] ] which is plotted in fig .
[ fig.1 ] .
the maximum occurs at @xmath42 , the maximum grooming , with the value @xmath43 .
this indicates that the @xmath2 is quite repulsive at this setting .
of course for @xmath27 or @xmath44 , the total energy is zero .
this shows that , even for the artificial case of zero separation , the @xmath2 energy is very dependent on grooming .
we now study , in the product ansatz , the energy of the @xmath2 system as a function of the separation and relative grooming .
this can only be done numerically .
we now use the full lagrangian including the finite pion mass term .
we put the @xmath11 and @xmath12 on a 3d lattice with the two solitons a distance @xmath45 apart on the @xmath46 axis .
the product ansatz is @xmath47 where @xmath48 is the @xmath1 field for a single skyrmion and @xmath22 is the grooming matrix . in skyrme units for length ( @xmath49 fm ) ,
the spatial extension of lattice we use is @xmath50 .
we evaluate the derivatives of the @xmath14-field by the two - point difference : @xmath51.\ ] ] with @xmath52 points on the lattice and @xmath53 , we find that , for large separations , the total energy is within one percent of twice the single skyrmion mass .
we calculate the energy for three interesting configurations : 1 .
no grooming ( @xmath54 ) 2 . relative rotation of @xmath9 around @xmath46-axis ( @xmath46-@xmath9 ) 3 . relative rotation of @xmath9 around @xmath55-axis ( @xmath55-@xmath9 ) the results
are shown in fig .
[ fig.2 ] . recall that the separation is along the @xmath46 axis .
it is instructive to compare the energy of the @xmath2 system with the corresponding result for the ss system @xcite .
the @xmath55-@xmath9 grooming is the most attractive configuration for @xmath6 , while it is the most repulsive for @xmath2 .
the @xmath46-@xmath9 grooming , while being the most repulsive for @xmath6 , is mildly attractive for @xmath2 .
finally the ungroomed case is the most attractive for @xmath2 while it is mildly repulsive for @xmath6 .
this also leads to speculation about the speed the speed of annihilation . in the calculation of @xcite ,
the starting configuration is ungroomed and annihilation happens very fast .
this may reflect the fact that this is the the most attractive channel .
the physical nucleon is a linear superposition of groomed skyrmions and perhaps in this case annihilation will proceed slower than that seen in @xcite .
we now turn to an expansion of the energy in the relative grooming variables as a first step in obtaining the projection of the @xmath2 interaction onto the @xmath3 interaction .
we follow the methods developed for obtaining the @xmath4 interaction from the @xmath6 .
as in the calculation of @xcite for @xmath6 , the energy for @xmath2 can be expanded in the variables @xmath56 and @xmath57 , with the relative grooming matrix @xmath58 and @xmath59 the vector connecting the centers of the two solitons . for the @xmath6 , the full expansion is @xmath60 where @xmath61 , @xmath62 are functions of @xmath45 . for @xmath2 , the symmetry of @xmath63
is broken by the product ansatz and we need three additional terms for a consistent expansion @xmath64 these terms odd in @xmath59 are an artifact of the asymmetry of the product ansatz and should be discarded .
one can use the symmetrized energy @xmath65 to extract @xmath66 to @xmath67 , since the @xmath68 to @xmath69 terms drop out in this combination .
the six terms in ( 11 ) can be expressed in terms of operators in the baryon space using the algebraic methods introduced in @xcite .
one quantizes each skyrmion with a @xmath70 algebra and then the relevant operators and baryon states are easily constructed in terms of the operators of those algebras .
the method was developed in @xcite for the @xmath4 system , but since each skyrmion gets its own algebra , the method can be taken over without alteration to the @xmath3 system .
the @xmath6 or @xmath2 interaction can be expanded in terms of three operators , the identity and the operators @xmath71 and @xmath72 given by @xmath73 /n_c^2.\end{aligned}\ ] ] here @xmath74 and @xmath24 label the two different set of bosons , used to realize the @xmath70 algebras .
@xmath75 is a one - body operator with spin 1 and isospin 1 .
the semiclassical ( large-@xmath0 ) limit of these operators can be given in terms of @xmath59 and @xmath22 as @xcite @xmath76 we can therefore expand the interaction in @xmath71 and @xmath72 , as an alternative to eq .
( [ eq.vexpand1 ] ) .
@xmath77 the relations between @xmath61 and @xmath78 can be found in eq .
( 24 ) of @xcite .
the advantage of the algebraic method is it allows us to study both the large @xmath0 limit , and to include finite @xmath0 effects explicitly in a systematic way .
it also makes taking baryon matrix elements quite easy . as in the @xmath6 case ,
we find for @xmath2 that the terms quadratic in @xmath72 and @xmath71 are quite small , and so we neglect them . hence we can write @xmath79 we can use the algebraic methods of @xcite to take the @xmath3 matrix element of our interaction . keeping only the leading terms
we find @xmath80 ( \tau_1\cdot\tau_2 ) \label{eq : v_nuconly}\ ] ] with @xmath81 here @xmath82 is a finite @xmath0 correction factor , @xmath83 .
this gives the nucleons only projection of the @xmath2 interaction . to obtain the full phenomenological interaction it is necessary to include the effects of @xmath7 and @xmath8 admixtures that become important as the baryons approach each other
we now turn to those admixtures .
the @xmath3 potential in eq .
( [ eq : v_nuconly ] ) is calculated by projecting eq .
( [ eq : potential_wz ] ) to the nucleon degrees of freedom only .
this is certainly the correct procedure for large separation .
however as the skyrmion and antiskyrmion approach , they can deform . in terms of the baryon degrees of freedom and @xmath84 that means excitation of the @xmath7 and @xmath8 intermediate states .
all that is required to define a @xmath3 interaction is that the particles be @xmath3 asymptotically . they may deform or excite as they wish as they interact . in the @xmath4 case we saw
that this intermediate excitation plays a significant role in the intermediate range attraction @xcite . for neutral atoms , a corresponding virtual excitation process leads to the attractive van der waals force at large distance . for skyrmions , beside this
state mixing , there is a dynamical distortion that goes beyond the product ansatz .
this is a crucial part of the @xmath4 interaction , but the corresponding @xmath2 distortion is beyond the scope of this paper . for the @xmath4 system
this distortion , coupled with the state mixing , is crucial for getting the mid - range attraction .
for the nucleon- antinucleon system we expect similar enhancements of attraction coming from the distortion .
the effects of distortion and state mixing both come in at distances where the product ansatz can be expected to fail . thus our short and mid - range results with the product ansatz , even including state mixing ,
should be taken as only indicative and not as the final word .
as in the @xmath4 system , we first study the effect of mixing @xmath7 and @xmath8 on the energy perturbatively and then use the born - oppenheimer method to consider the effect exactly in the limited subspace .
we first include the effects of the intermediate states , @xmath85 , @xmath86 and @xmath87 on the nucleon antinucleon potential perturbatively .
since we are using separate @xmath70 algebras for each skyrmion , the results for @xmath4 in @xcite can be carried over to the @xmath3 problem , with eq .
( 15 ) in @xcite being the perturbation correction .
@xmath88(v_2 ^ 2 + 2v_3 ^ 2)\right .
\nonumber \\ & & + ( \mbox{\boldmath $ \sigma$}^1\cdot \mbox{\boldmath $ \sigma$}^2)\left [ -\frac{1}{18}q_n^2p_0^\tau+(\frac{16}{81}p_n^2-\frac{5}{162}q_n^2)p_1^\tau \right](v_2 ^ 2-v_3 ^ 2)\\ & & \left .
+ ( 3\mbox{\boldmath $ \sigma$}^1\cdot\hat{\mbox{\boldmath $ r$ } } \mbox{\boldmath $ \sigma$}^2\cdot\hat{\mbox{\boldmath $ r$ } } -\mbox{\boldmath $ \sigma$}^1\cdot\mbox{\boldmath $ \sigma$}^2 ) \left[-\frac{1}{18}q_n^2p_0^\tau+(\frac{16}{81}p_n^2 -\frac{5}{162}q_n^2)p_1^\tau \right](v_2 ^ 2-v_2 v_3 ) \right\}. \nonumber\end{aligned}\ ] ] here @xmath89 is the @xmath10-@xmath7 mass difference , @xmath90 is a projection operator onto isospin @xmath75 , and @xmath91 is another finite @xmath0 correction factor with the value @xmath92 .
this expresses the leading order correction from state mixing to the @xmath3 interaction of eq .
( [ eq : v_nuconly ] ) in terms of the @xmath2 terms of eq .
( [ eq : potential_wz ] ) . recall that unlike the work in @xcite we are here using the product ansatz to calculate the @xmath2 interaction rather than a full dynamical scheme .
we now turn to a full diagonalization of the interaction in the @xmath3 , @xmath85 , @xmath93 and @xmath87 space .
this is the born - oppenheimer approximation , and it is valid for @xmath84 .
there are three energy scales or time scales in the problem .
the fastest or highest energy scale comes in rearrangements of the pion field itself .
these we are modeling using the product ansatz and correspond to energies on the scale of baryon masses .
the intermediate scale is set by the @xmath10 @xmath7 energy difference .
this is an order @xmath94 effect .
finally the @xmath3 interaction is the smallest energy scale and it is determined by the matrix diagonalization .
we first need the matrix element of the potential eq .
( [ eq : potential_wz ] ) in the space of @xmath3 , @xmath85 , @xmath93 and @xmath87 in the angular momentum coupled form .
this has been calculated in eq .
( 22 ) of @xcite for the baryon - baryon case and the formula remains valid for baryon - antibaryons .
@xmath95 the relevant reduced matrix elements are @xmath96 the matrix element of the kinetic part are taken to be very simple , @xmath97+\frac{l(l+1)}{2m_{i_1i_2}r^2}\right).\ ] ] here @xmath98 is the reduced mass , @xmath99 , with @xmath100 mev and @xmath101 mev .
the mass difference is @xmath102 . for the purpose of comparison , we parametrize the full @xmath3 interaction by @xmath103 the potentials have explicit isospin dependence due to the mixing with states of @xmath7 . to determine the adiabatic potential for @xmath3 , we start at large @xmath45 where we have nucleons only . as we move to smaller distance , we diagonalize the @xmath104 matrix and follow continuously the eigenvalue corresponding to the @xmath3 channel .
we then subtract the expectation value of @xmath105 to obtain the adiabatic interaction energy as a function of @xmath45 .
we first consider the case @xmath106 .
for @xmath107 ( note that nucleon and antinucleon have opposite intrinsic parity ) , we have three channels @xmath108 , @xmath109 and @xmath110 .
the lowest eigenvalue of the hamiltonian @xmath104 should be identified with @xmath111 for @xmath112 , there are three channels @xmath113 , @xmath114 and @xmath115 .
the lowest eigenvalue should be equated to @xmath116 we consider one more set of states with @xmath117 and there are six channels : @xmath118 , @xmath119 , @xmath120 , @xmath121 , @xmath122 and @xmath123 .
the matrix element to identify the lowest eigenvalue with is @xmath124 from these three linear combinations of @xmath125 , @xmath126 and @xmath127 in eq .
( [ eq : v01 ] ) to eq .
( [ eq : v03 ] ) , the potentials in eq .
( [ vpara ] ) are easily solved for @xmath106 .
a similar calculation applies for @xmath128 , except that now @xmath85 and @xmath93 channels also appear .
for each total isospin @xmath129 , we calculate @xmath130 and @xmath131 in eq .
( [ vpara ] ) as outlined in the last section .
we then compare these result with the phenomenological potentials of brian - phillips @xcite and of the nijmegen @xcite group .
these are potentials based on meson exchange at large distances and phenomenology , including an absorptive part to model annihilation , at small distances .
the meson exchange part of these potentials for @xmath3 is obtained from the corresponding @xmath4 potentials by @xmath5-parity transform the contribution of a particular meson for @xmath3 is equal to its part in @xmath132 multiplied by the meson s @xmath5-parity .
we only compare with the scalar , tensor and spin - spin parts of the potentials .
the spin - orbit force is of higher order in @xmath94 and we have not calculated it in the skyrme picture .
we should note that various cutoffs are used in the brian - phillips , nijmegen and other similar potentials . as a result at distance 1 fm or less the strength of the potentials
can be significantly different from their meson exchange value .
in addition , at distance less than 1 fm , the interaction is dominated by the absorptive potential of order 1 gev .
furthermore , at these short distances the entire static skyrme approach , to say nothing of the product ansatz , is no longer meaningful .
hence we should not place any faith on comparisons of our results with the phenomenological potentials around 1 fm or less . at intermediate distances , between 1 fm and 2 fm
, the results from our skyrme approach to the @xmath4 interaction suggest that the product ansatz is on the right track , but that careful comparison with phenomenological potentials requires a more complete calculation of the skyrme dynamics . in particular we expect the product ansatz to underestimate mid - range attraction , as it does for @xmath4 .
this is basically a consequence of the variational theorem . with these thoughts in mind
, let us turn to our results .
figure 3 shows our results for the @xmath106 part of the central potential .
we see that the effects of @xmath7 mixing either in perturbation theory or full born - oppenheimer diagonalization is significant , but still does not begin to agree with the strong central , mid - range attraction seen in the phenomenological potentials .
this is the fault of the product ansatz we referred to above
. it will be important to see if complete skyrme calculations can repair this fault .
figure 4 shows the @xmath128 central potential . here
the effects of @xmath7 mixing are more striking since for @xmath128 single @xmath7 intermediate states are permitted .
now we do find some central attraction , but not as much as is seen phenomenologically .
note that where they differ , the full diagonalization result has superior credentials to the perturbation theory result , and it is the diagonalization result that is too weak .
again we must await full skyrme calculation in this channel .
figure 5 shows the @xmath106 spin - spin part of the interaction . except at the smallest distances ,
the results are very satisfactory .
note that this is remarkable , since the spin - spin interaction is very weak ( note the scale in figure 5 ) , and hence arises from cancellation of much larger terms .
we believe it is significant that the skyrme picture can reproduce this scale and even the correct sign .
the effects of cancellations are even more striking in figure 6 which shows the @xmath128 spin - spin interaction . again
the order of magnitudes are correct while the sign depends sensitively on how we do the @xmath7 mixing .
note that the phenomenological potentials are consistent with zero .
finally in figures 7 and 8 we show the @xmath106 and @xmath128 tensor potentials .
all calculations of these agree since they are dominated by one - pion exchange .
( recall that the skyrme picture in the product ansatz gets one pion exchange right . )
hence except for the central attraction , the product ansatz gives a credible account of the nucleon - antinucleon potential , and we understand how the product ansatz fails for the central attraction .
note that there are no free parameters on our calculation .
we have shown that the skyrme picture with the product ansatz is a reasonable first step to obtaining the real part of the nucleon - antinucleon interaction .
we also understand how doing the skyrme dynamics better can repair the lack of central attraction we find here . hence the next step is to do that dynamics . then combing this picture of nucleon - antinucleon interactions in the entrance channel based on the skyrme model with our previous work on annihilation channels described by this model , we hope to have a unified picture of annihilation based on the large @xmath0 , qcd inspired skyrme picture .
this work was supported in part by a grant from the national science foundation . | we calculate the nucleon - antinucleon static potential in the skyrme model using the product ansatz and including some finite @xmath0 ( number of colors ) corrections . the mid and long range part of the spin - spin and tensor force are correctly given in both iso - spin channels while the central interaction has insufficient mid - range attraction .
this is a well known problem of the product ansatz that should be repaired with better skyrme dynamics . |
the radiation from luminous quasi stellar objects ( qsos ) can currently be studied up to redshift 6.5 and thus allows one to probe , over much of the universe s age , the properties of their central black holes ( bhs , e.g. , * ? ? ?
* ) , their broad line region ( blr , e.g. , * ? ? ?
* ) , and the ionization state of intergalactic gas ( e.g. , * ? ? ?
the most distant qsos are of importance for both the study of galaxy evolution and large scale structure . among the most surprising discoveries in this field
are the detection of bh masses
@xmath9 m@xmath4 ( e.g. , * ? ? ?
* ; * ? ? ?
* ) and the apparent constancy of the observed qso blr properties with redshift , even up to @xmath10 ( e.g. , * ? ? ?
* ) , corresponding to an age of 870myr .
e.g. , the / ( * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* hereafter k07 ) and / @xcite emission line ratios observed in the blr of @xmath11 qsos are as high as that observed in low - redshift objects .
thus , near - solar metallicity in the centers of some of the earliest systems and accretion of @xmath9 solar masses in a central bh can be attained in less than 1gyr .
one caveat , however , has to be taken into account : until now only the most luminous objects at @xmath8 have been studied .
these objects may experience faster evolution than less luminous objects and therefore resemble low to intermediate redshift qsos already at @xmath2 .
secondly , the relationships established using low - luminosity agn ( at low redshift ) may not be valid for objects at the upper end of the luminosity distribution .
as an example , the determination of the bh mass in active galaxies was calibrated with reverberation mapping observations of agn that are one to two orders of magnitude fainter than the @xmath2 qsos observed until now @xcite .
it is therefore important to extend @xmath12 studies to lower luminosity objects . here
we present the first observations of a qso at @xmath13 with luminosity lower than l@xmath14 ergs@xmath15 : the qso ( hereafter j0303 - 0019 ) .
this qso was detected in the sloan digital sky survey ( sdss ) deep stripe with a magnitude of @xmath16 .
its discovery spectrum shows a very strong and narrow ly@xmath17 emission line at @xmath18 @xcite .
as this line falls in the @xmath19-band , the actual continuum luminosity is low , @xmath20 , 1.5 magnitudes fainter than the mean of the @xmath8 qso samples studied in k07 and @xcite .
in addition , this is one of the two out of about twenty @xmath2 qsos that are not detected at 24@xmath6 m with mips / spitzer , implying a rest frame 3.5@xmath6m/4400ratio far lower than for any other known qso ( * ? ? ? * l. jiang et al . in prep . ) .
nir spectroscopy of j0303 - 0019 in the @xmath21 and @xmath22 bands was obtained in service mode on december 24 and november 26 and 28 , 2007 , respectively , with isaac @xcite at vlt ut1 .
we used the short wavelength arm , equipped with a 1024@xmath231024 hawaii rockwell array which has a pixel scale of 0.147 .
the 1 slit results in a spectral resolution of 550 and 450 for the @xmath21 and @xmath22 bands , respectively .
observations were carried out in a standard abba offset pattern , with additional dithering to reduce the influence of bad pixels .
individual exposures were 148s , repeated 16 times per observing block .
nine such blocks were obtained in @xmath22 band , for a total exposure time of @xmath245.9hr , and two in @xmath21 band , for a total time of @xmath241.3hr .
the data were reduced using eso s isaac pipeline software , which does flat fielding , background subtraction , rectification and combination of the individual frames in one observing block .
we used sky lines to determine the wavelength calibration , obtaining an absolute 1@xmath25 error of 3 and 7 , respectively , in @xmath21 and @xmath22 band .
one - dimensional spectra were extracted from the pipeline products in iraf .
b and g stars with known nir photometry were used to correct for telluric absorption and initial flux calibration of the one - dimensional spectra .
finally , the one - dimensional spectra were averaged to obtain the final spectra .
these were flux calibrated using the broad band magnitudes obtained with the calar alto 3.5 m telescope ( see sec.[ssec : nir_img ] ) .
the reduction is similar to that carried out by k07 , where more details are reported . for accurate flux calibration ,
the qso sdss j0303 - 0019 was imaged with the prime - focus nir camera omega2000 @xcite at the 3.5 m telescope at calar alto , spain .
this instrument has a 2k@xmath232k hawaii-2 detector with a pixel scale of 0.45 , resulting in a field of view of [email protected] .
j0303 - 0019 was observed on september 10 and 11 , 2008 , for 20 , 20 , 25 , and 77 minutes in @xmath26 , and @xmath27 bands , respectively .
after dark subtraction and flat fielding with twilight exposures , the images were sky subtracted , registered and averaged with xdimsum in iraf .
the seeing full width at half maximum ( fwhm ) on the combined images is 1.3 , 1.4 , 1.3 , and 1.0 , respectively .
the combined images were astrometrized using the public service provided by astrometry.net @xcite using a second order polynomial to correct for distortions ( which are small for omega2000 ) .
subsequently , we used about 80 stars from the 2mass all - sky catalog of point sources @xcite to obtain the magnitude zero points . for @xmath21band , we used the ukidss second data release @xcite catalog .
only half of the observed field is covered by ukidss , but this still provides 31 stars usable for the flux calibration .
systematic errors in the flux calibration are estimated to be below 0.1mag .
the resulting fluxes ( using mag_best in sextractor , * ? ? ?
* ) are reported in table [ tbl : nir_mags ] .
rrrr & & & + [email protected] & [email protected] & [email protected] & [email protected] +
rrrrrrrrrr & @xmath17 & & & & & & & + & & & & & & & & & + ly@xmath17 ( broad ) & @xmath29 & [email protected] & 8599@xmath281 & 125@xmath305 & 4350@xmath28165 & 255@xmath306 & 87@xmath282 + ly@xmath17 ( narrow ) & & & & 29@xmath302 & 1000@xmath3070 & 109@xmath2811 & 37@xmath284 + & & [email protected] & 8758@xmath281 & 81@xmath302 & 2760@xmath3080 & 90@xmath302 & 32@xmath281 + & @xmath29 & [email protected]&10953@xmath281 & 65@xmath302 & 1780@xmath3065 & 54@xmath302 & 37@xmath281 & [email protected] & [email protected] + & @xmath31 & [email protected]&10953@xmath281 & 65@xmath303 & 1772@xmath3070 & 54@xmath302 & 36@xmath281 & [email protected] & [email protected] + & @xmath32 & [email protected]&10953@xmath281 & 64@xmath303 & 1760@xmath3070 & 53@xmath302 & 35@xmath281 & [email protected] & [email protected] + & @xmath29 & [email protected]&19804@xmath284&149@xmath28 9 & 2261@xmath28130 & 16@xmath301 & 25@xmath281 & [email protected] & [email protected] + & @xmath31 & [email protected]&19803@xmath284&154@xmath28 8 & 2340@xmath28125 & 17@xmath301 & 27@xmath281 & [email protected] & [email protected] + & @xmath32 & [email protected]&19801@xmath283&167@xmath28 8 & 2525@xmath28120 & 19@xmath301 & 32@xmath281 & [email protected] & [email protected] + in the top panel of fig .
[ fig : spectra ] , we show a close - up version of the optical spectrum , first presented in @xcite , taken at the w.m . keck observatory . for comparison ,
the sdss qso composite spectrum published by ( * ? ? ?
* dashed curve ) , normalized to emission line free regions of our observed spectra , is overlayed , showing that the emission lines of j0303 - 0019 are much stronger and narrower than those present in the composite spectrum .
we have made a simultaneous fit of the broad and narrow ly@xmath17 and lines in the optical spectrum , after subtracting a standard @xcite power law continuum with a slope @xmath33 as in @xcite .
we fit only the spectrum redwards of the strong absorption feature seen in the ly@xmath17 line .
the best fit parameters are listed in table [ tbl : line_meas ] , and the best fit curves are over plotted in fig . [
fig : spectra ] . the equivalent width
( ew ) and fwhm of measured by us are about 20% higher than those reported by @xcite , who fixed the central wavelength of all fitted lines to the same redshift . in the middle and bottom panels of fig .
[ fig : spectra ] , we show the vlt isaac @xmath21- and @xmath22-band spectra , containing the and emission lines , respectively .
there is also a tentative detection of /]1398 at 9900 in the @xmath21-band , which we do not try to fit .
the obtained nir photometry of j0303 - 0019 ( table [ tbl : nir_mags ] ) is consistent with a power law sed with slope @xmath34 , as measured in lower redshift qsos @xcite .
imperfect flux calibration of the nir spectra , however , may affect the slope of the reduced @xmath21 and @xmath22 band spectra .
indeed , the slope observed in the @xmath22 band seems redder and may be fit better by a slope of @xmath35 or even @xmath36 . as the nir spectra are too short and the data not of sufficient quality to constrain the slope , we carry out emission line fits employing a range of slopes @xmath37 fixed to = 0 , -1.0 , and -1.5 , to cover the complete range of plausible slopes .
we take the slope of @xmath35 as the best compromise and use the parameters derived from fits employing the two extreme values to construct conservative error estimates on the fitted parameters .
the and lines are fit by single gaussian functions on top of the co - addition of the power - law continuum , a balmer pseudo - continuum and a template of lines ( see also * ? ? ?
the strength of the latter two components were scaled with the power - law continuum , according to their proportions in the composite qso spectrum by @xcite .
as the strength of the emission is not well constrained by the current spectra , but clearly lower than four times the strength in the sdss composite spectrum , we have also carried out the fitting assuming a four times stronger and four times weaker contribution to establish proper error estimates .
the resulting best - fit parameters are displayed in table [ tbl : line_meas ] .
the and lines have ( rest frame ) fwhm@xmath38 less than 2000 and 2600 kms@xmath15 , and ew@xmath38 of 36@xmath281 and 27@xmath39 , respectively .
the mean fwhm@xmath38 of these lines in the sdss qso sample of ( * ? ? ? * hereafter s08 ) is 5700 and 5850 kms@xmath15 , respectively .
only 2% and 1% of the sdss qsos have and fwhm@xmath38 smaller than 2000 and 2600 kms@xmath15 , respectively .
the widths of the and lines , combined with an independent estimate of their characteristic distance from the center , can be used as a measure of the mass of the black hole ( m@xmath40 ) .
the relation between line width , luminosity and m@xmath40 has been calibrated empirically at low redshift using reverberation mapping ( e.g. , * ? ? ?
as in k07 and @xcite , we use the relations for and presented by @xcite and @xcite , respectively , which are also those applied to a sample of more than 60,000 sdss qsos by s08 . due to the scatter of individual data points around this relation m@xmath40 can be determined only within a factor of three . for j0303 - 0019
, we derive m@xmath40 with these methods and take into account the range of power - law slopes and template strength described in sec .
[ ssec : spectra ] .
we obtain m@xmath40 of 1.6@xmath41 and 2.5@xmath4210@xmath43 m@xmath4 , respectively , from the and lines .
these are the lowest m@xmath40 derived in a @xmath13 qso to date , showing that , with this sample of fainter @xmath8 qsos , we are indeed probing down the mass function of bhs .
we note that the central wavelength of the line is blue shifted by 350 kms@xmath15 w.r.t . the line .
according to s08 , this modest shift ( @xmath44 kms@xmath15 ) implies that the line profile is dominated by gravitational effects and that m@xmath40 can therefore be reliably estimated from this line .
the m@xmath40 derived from the line widths and corresponding eddington luminosities of @xmath45 = 46.31@xmath46 , and 46.51@xmath47 ergs@xmath15 , respectively , are small compared to the qso s bolometric luminosity of @[email protected] ergss@xmath15 , implying a high eddingtion ratio ( r@xmath49 ) . from optical , ir , and radio data ( l. jiang et al , in prep . ) , the bolometric luminosity of j0303 - 0019 is determined as in @xcite to be @xmath50 ergss@xmath15 .
it is close to the values of @xmath51 = 46.62 , and 46.65 ergs@xmath15 obtained using the monochromatic luminosities at 1350 and 3000 , respectively , and the bolometric correction factors employed by s08 ( bc@xmath52 = 3.81 , bc@xmath53 = 5.15 ) . reconciling these values
, we obtain the estimated bolometric luminosity stated above .
the eddington ratio of most qsos is between 0.01 and 1 and only 804 out of the 62,185 sdss with m@xmath40 determined from their line width have r@xmath49@xmath54 ( s08 ) . the eddington ratios for j0303 - 0019 derived from the and lines are r@xmath49 = 1.9@xmath55 and 1.2@xmath56 ( see table [ tbl : line_meas ] and fig . [
fig : mass_lum ] ) .
less than 2% and 0.2% of the sdss qsos have such an r@xmath49()@xmath571.9 and r@xmath49()@xmath571.2 .
even if we assume that the other four ( faint ) @xmath2 qsos of the complete sample of qsos discovered in the sdss deep stripe @xcite have r@xmath49 @xmath581.2 , the fraction of l@xmath59 ergs@xmath15 qsos at @xmath2 with r@xmath49 @xmath571.2 is 33% , almost 40 times higher than at lower redshifts .
the chemical composition of gas clouds in the blr can be explored by comparing observed emission line ratios to models .
the heavy elements in the blr are products of the star formation history in the host galaxy and therefore indicators of past star formation in at least the very center of the host .
@xcite present several emission line ratios observed in sdss qsos at @xmath60 and models explaining the ratios .
according to @xcite , the strength of the / ratio is strongly correlated with the luminosity of the qso , which can be interpreted in terms of higher gas metallicity in more luminous qsos .
although j0303 - 0019 is one of the faintest qsos at @xmath2 , it has a @xmath61 of -26.2 , which makes it moderately luminous in comparison with the sdss qsos studied by @xcite .
we measure a ratio of / = [email protected] , the highest ratio among the known @xmath8 qsos , and also higher than those measured in the composite spectra of quasars in the luminosity range @xmath62 by @xcite .
this suggests that the blr has a high metal abundance , possibly as high as 10z@xmath4 ( see fig . 29 in
* ) , but the authors warn that estimates of the blr metallicity using only emission line ratios involving might be quite uncertain ( see also * ? ? ?
the intense uv radiation of the qso keeps a region around the qso ionized , the strmgren sphere , which will actually only be spherical if the igm surrounding the qso is homogeneous . outside this region ,
uv photons are absorbed by neutral hydrogen and this causes complete absorption blueward of the wavelength of redshifted ly@xmath17 , the so - called gp trough @xcite . for reasons discussed in @xcite ,
the strmgren sphere radius is difficult to measure and we rather use the size of the proximity zone ( r@xmath63 ) , defined as the region in which the transmitted flux is @xmath5710% of the ( extrapolated ) continuum emission when smoothed to a resolution of 20 ( as shown by the grey line in fig .
[ fig : spectra ] ) .
we measure the blue edge of this region to be at 8513@xmath281 ( dashed line in fig .
[ fig : spectra ] ) . combined with the systemic redshift of @xmath64 derived from the line ( redmost dashed line in fig .
[ fig : spectra ] ) , we compute ( k07 ) the co moving radius of the strmgren sphere to be r@xmath65 = [email protected] .
this is a typical value for luminous @xmath2 qsos ( see , e.g. , k07 , * ? ? ?
the average size of qso strmgren spheres , however , decreases by a factor @xmath242.5 over @xmath66 , consistent with an increase of the neutral fraction with redshift @xcite .
sdss j0303 - 0019 has the lowest bh mass known at @xmath2 and is a truly peculiar qso : it has one of the highest eddington ratios observed for a qso , it has the highest / ratio observed in @xmath2 qsos , which is also significantly higher than that observed in lower redshift composite spectra of qsos , and it is one of the two @xmath2 qsos that is not detected in deep 24@xmath6 m observations ( * ? ? ?
* l. jiang et al . in prep . ) .
the other qso which is not detected at 24@xmath6 m is ( hereafter j0005 - 0006 ) , which also has a high eddington ratio : 2.4 as derived from the based m@xmath40 and 1.0 as derived from the based m@xmath40(k07 ) .
detections at 250ghz and 1.4ghz have also been attempted but neither of the objects were detected down to rms values of 0.51 , 0.48mjy and 62 , 130@xmath6jy , respectively @xcite .
in contrast , these qsos were detected at 3.6 , 4.5 and 5.8@xmath6 m with irac / spitzer ( * ? ? ? * l. jiang et al . in prep . ) , implying a rest frame 3.5@xmath6m/4400 ratio far lower than for any object in the @xmath2 qso sample , and lower than for any of the pg qsos observed by @xcite and @xcite for which the sensitivity at 3.5mu rest - frame is not an issue .
@xcite show that the optical / ir seds of the observed @xmath2 qsos can be fit by two components : a power - law representing an accretion disk and a black - body representing hot dust .
the respective ( 1@xmath25 ) upper limits of 21@xmath6jy and 9@xmath6jy at 24@xmath6 m for j0303 - 0019 and j0005 - 0006 , significantly lower than the expected flux of @xmath24160@xmath6jy based on seds of other @xmath2 qsos , suggests the absence of hot dust . the implied lack of dust
could be explained if the gas is metal poor , as in a galaxy host which has not experienced major star formation yet .
however , this interpretation would contradict the potentially high metallicity based on the measured / ratio . alternatively
, the non - detection of the hot dust component could be due to different dust properties , or due to a surrounding cool dust torus hiding the hot component @xcite .
we conclude that , with the observation of qsos with bolometric luminosity @xmath67 ergs@xmath15 , we start to probe a population of qsos that shows different observed properties as compared to higher luminosity @xmath2 qsos and lower
redshift qsos of all luminosities .
we can only speculate that the hosts of the lower
luminosity qsos at @xmath2 may be younger than higher
luminosity qsos at the same redshift because the host galaxies of the latter are more massive and have therefore formed earlier .
it is , however , unclear at this stage what physical mechanisms are causing the observed differences .
we thank the anonymous referee for comments that helped to improve the paper . based on observations carried out at the european southern observatory , chile ( program no .
080.a-0794 ) and at the calar alto observatory , spain .
jk acknowledges financial support from _ deutsche forschungsgemeinschaft _ ( dfg ) grant sfb 439 .
xf and lj acknowledge supports from nsf grants ast 03 - 07384 and ast 08 - 06861 and a packard fellowship for science and engineering .
xf acknowledges additional support from the max planck society and a guggenheim fellowship .
dr acknowledges support from nasa through hubble fellowship grant hst - hf-01212.01a awarded by the stsci , operated by aura , under contract nas 5 - 26555 .
this publication makes use of data products from the two micron all sky survey , which is a joint project of the university of massachusetts and the infrared processing and analysis center / california institute of technology , funded by the national aeronautics and space administration and the national science foundation .
_ facilities : _ , | we present sensitive near infrared vlt isaac spectroscopic observations of the @xmath0 quasar sdss [email protected] . this qso is more than a magnitude fainter than other qsos at @xmath2 for which nir spectroscopy has been obtained to date and is therefore presumably more representative of the qso population at the end of cosmic reionization . combining rest frame uv continuum luminosity with the width measurements of the and lines , we derive a black hole mass of @xmath3 m@xmath4 , the lowest mass observed for @xmath2 qsos to date , and
derive an eddington ratio of 1.6@xmath5 , amongst the highest value derived for qsos at any redshift .
the spitzer 24@xmath6 m non detection of this qso does not leave space for a significant hot dust component in its optical / near infrared sed , in common with one other faint qso at @xmath7 , but in contrast to more than twenty more @xmath7 qsos and all known lower redshift qsos with sufficiently deep multi - wavelength photometry .
we conclude that we have found evidence for differences in the intrinsic properties of at least one @xmath8 qso as compared to the lower redshift population . |
the aim of this paper is to define and provide sources of examples of stable anti - yetter - drinfeld modules .
they play the role of coefficients for hopf - cyclic theory @xcite .
in particular , we claim that modular pairs in involution of connes and moscovici are precisely 1-dimensional stable anti - yetter - drinfeld modules . throughout the paper
we assume that @xmath0 is a hopf algebra with a bijective antipode . on the one hand ,
the bijectivity of the antipode is implied by the existence of a modular pair in involution , so that then it need not be assumed . on the other hand , some parts of arguments might work even if the antipode is not bijective .
we avoid such discussions .
the coproduct , counit and antipode of @xmath0 are denoted by @xmath1 , @xmath2 and @xmath3 , respectively . for the coproduct we use the notation @xmath4 , for a left coaction on @xmath5
we write @xmath6 , and for a right coaction @xmath7 .
the summation symbol is suppressed everywhere .
we assume all algebras to be associative , unital and over the same ground field @xmath8 .
the symbol @xmath9 stands for the algebra of polynomial functions on @xmath10 .
p.m.h . acknowledges the marie curie fellowship hpmf - ct-2000 - 00523 and kbn grant 2 p03a 013 24 .
all authors are grateful to t. brzeziski and m. furuta for their help and comments , and to d. perrot for the french translation .
it turns out that , in order to incorporate coefficients into cyclic theory , we need to alter the concept of a yetter - drinfeld module by replacing the antipode by its inverse in the yetter - drinfeld compatibility condition between actions and coactions .
we call the modules - comodules satisfying the thus modified yetter - drinfeld compatibility condition _ anti - yetter - drinfeld modules_. just as yetter - drinfeld modules come in 4 different versions depending on the side of actions and coactions ( see @xcite for a general formulation ) , so do the anti - yetter - drinfeld modules .
all versions are completely equivalent and can be derived from one another by replacing a hopf algebra @xmath0 by @xmath11 , @xmath12 , or @xmath13 , respectively .
+ [ ayd ] let @xmath0 be a hopf algebra with a bijective antipode @xmath3 , and @xmath5 a module and comodule over @xmath0 .
we call @xmath5 an anti - yetter - drinfeld module iff the action and coaction are compatible in the following sense : @xmath14 to make cyclic theory work , we also need to assume that the action splits coaction , i.e. , for all @xmath15 , @xmath16 , @xmath17 , @xmath18 , @xmath19 , for the left - left , left - right , right - left , and right - right versions , respectively .
we call modules satisfying this condition _ stable_. let us emphasize that it is the anti - yetter - drinfeld condition rather than the yetter - drinfeld condition that makes the homomorphism @xmath20 @xmath0-linear and @xmath0-colinear
. therefore the stability condition @xmath21 suits the former and not the latter .
the first class of examples of stable anti - yetter - drinfeld modules is provided by modular pairs in involution @xcite . since such pairs occur naturally in different contexts , lemma [ 1dim ] and lemma [ affine ] guarantee ample amount of examples of anti - yetter - drinfeld modules .
+ [ 1dim ] let the ground field @xmath8 be a right module over @xmath0 via a character @xmath22 and a left comodule over @xmath0 via a group - like @xmath23 . then @xmath24 is a stable anti - yetter - drinfeld module _ if and only if _ @xmath25 is a modular pair in involution .
the anti - yetter - drinfeld modules do not form a monoidal category themselves , but rather a so - called @xmath26-category over the category of yetter - drinfeld modules ( see @xcite for details ) . more precisely
: + [ affine ] let @xmath27 be a yetter - drinfeld module and @xmath5 an anti - yetter - drinfeld module .
then @xmath28 is an anti - yetter - drinfeld module via @xmath29 , @xmath30 , for the left - left case , and via @xmath31 , @xmath32 , for the left - right case .
similarly , @xmath33 is an anti - yetter - drinfeld module via @xmath34 , @xmath35 , for the right - left case , and via @xmath36 , @xmath37 , for the left - right case .
note that , just as the right - right yetter - drinfeld modules are entwined modules @xcite for the entwining @xmath38 , the right - right anti - yetter - drinfeld modules are entwined with respect to @xmath39 .
( other cases are completely analogous . ) an intermediate step between modular pairs in involution and stable anti - yetter - drinfeld modules is given by matched and comatched pairs of @xcite .
whenever the antipode is equal to its inverse , the difference between the yetter - drinfeld and anti - yetter - drinfeld conditions disappears . for a group ring hopf algebra @xmath40 ,
a left @xmath0-comodule is simply a @xmath41-graded vector space @xmath42 , where the coaction is defined by @xmath43 .
an action of @xmath41 on @xmath5 defines an ( anti-)yetter - drinfeld module if and only if for all @xmath44 and @xmath45 we have @xmath46 .
the stability condition means simply that @xmath47 for all @xmath48 , @xmath45 .
a very concrete classical example of a stable ( anti-)yetter - drinfeld module is provided by the hopf fibration . then @xmath49 and @xmath50 . since @xmath51
, we have a natural left action of @xmath52 on @xmath53 .
its pull - back makes @xmath5 a left @xmath0-comodule . on the other hand
, one can view @xmath53 as the set of all traceless matrices of @xmath52 .
the pull - back of this embedding @xmath54 together with the multiplication in @xmath55 defines a left action of @xmath0 on @xmath5 .
it turns out that the equivariance property @xmath56 guarantees the anti - yetter - drinfeld condition , and this combined with the injectivity of @xmath57 ensures the stability of @xmath5 .
this stability mechanism can be generalized in the following way .
+ let @xmath5 be an algebra and a left @xmath0-comodule .
assume that @xmath58 is an epimorphism of algebras and the action @xmath59 makes @xmath5 an anti - yetter - drinfeld module .
assume also that @xmath60 .
then @xmath5 is a stable module .
another source of examples is provided by hopf - galois theory .
these examples are purely quantum in the sense that the employed actions are automatically trivial for commutative algebras .
to fix the notation and terminology , recall that an algebra and an @xmath0-comodule is called a comodule algebra if the coaction is an algebra homomorphism .
an @xmath0-extension @xmath61 is called hopf - galois iff the canonical map @xmath62 , @xmath63 , is bijective .
the bijectivity assumption allows us to define the translation map @xmath64 , @xmath65}\otimes_bh^{[2]}$ ] ( summation suppressed ) .
it can be shown that when everything is over a field ( our standing assumption ) , the centralizer @xmath66 of @xmath67 in @xmath68 is a subcomodule of @xmath68 .
on the other hand , the formula @xmath69}ph^{[2]}$ ] defines a right action on @xmath70 called the miyashita - ulbrich action .
this action and coaction satisfy the yetter - drinfeld compatibility condition ( * ? ? ? * ( 3.11 ) ) .
the following proposition modifies the miyashita - ulbrich action so as to obtain stable anti - yetter - drinfeld modules .
+ [ mu ] let @xmath71 be a hopf - galois @xmath0-extension such that @xmath67 is central in @xmath68 . then @xmath68 is a right - right stable anti - yetter - drinfeld module via the action @xmath72}p(s^{-1}(h))^{[1 ] } $ ] and the right coaction on @xmath68 .
the simplest examples are obtained for @xmath73 .
a broader class is given by the so - called galois objects @xcite
. then quantum - group coverings at roots of unity provide examples with central coinvariants bigger than the ground field ( see @xcite and examples therein ) .
finally , one can generalize proposition [ mu ] to arbitrary hopf - galois extensions by replacing @xmath68 by @xmath74 $ ] ( * ? ? ? * remark 4.2 ) .
for finite - dimensional hopf algebras , the yetter - drinfeld modules can be understood as modules over the drinfeld double @xcite .
much in the same way , the anti - yetter - drinfeld modules can also be understood as modules over a certain algebra .
this makes the usual notions and operations for modules , like projectivity or induction , directly available for anti - yetter - drinfeld modules . to this end
, the comodule structure of an anti - yetter - drinfeld module has to be converted into a module structure over the dual hopf algebra @xmath75 , so that from now on we assume that the hopf algebra @xmath0 is finite - dimensional .
+ note that the above product differs from the product in the drinfeld double of @xmath0 @xcite only by the additional squared antipode in the second factor .
to relate the modules over @xmath79 with anti - yetter - drinfeld modules , recall first that every right @xmath0-comodule @xmath5 becomes a left @xmath75-module via @xmath80 .
conversely , any left @xmath75-module yields a right @xmath0-comodule via @xmath81 . here
@xmath82 is a basis of @xmath0 and @xmath83 is the dual basis .
( of course , this comodule structure does not depend on the choice of a basis . ) using this , we get the following connection between the modules over @xmath79 and anti - yetter - drinfeld modules : + let @xmath0 be a finite - dimensional hopf algebra .
if @xmath5 is a left - right anti - yetter - drinfeld module , it becomes a left @xmath79-module by @xmath84 .
conversely , if @xmath5 is a left @xmath79-module , it becomes a left - right anti - yetter - drinfeld module by @xmath85 , @xmath86 .
here @xmath82 is a basis of @xmath0 and @xmath83 its dual basis .
the claim of lemma [ affine ] is reflected in the fact that although @xmath79 is not a hopf algebra itself , it can be shown that the formula @xmath87 makes @xmath79 a right comodule algebra over the drinfeld double @xmath88 .
dabrowski , l. ; hajac , p.m. ; siniscalco , p. : explicit hopf - galois description of @xmath90-induced frobenius homomorphisms .
enlarged proceedings of the isi guccia workshop on quantum groups , non commutative geometry and fundamental physical interactions , d. kastler , m. rosso , t. schucker ( eds . ) , commack new york , nova science pub , inc . , pp.279 - 298 , 1999 . | we define and study a class of entwined modules ( stable anti - yetter - drinfeld modules ) that serve as coefficients for the hopf - cyclic homology and cohomology . in particular , we explain their relationship with yetter - drinfeld modules and drinfeld doubles . among sources of examples of stable anti - yetter - drinfeld modules , we find hopf - galois extensions with a flipped version of the miyashita - ulbrich action . _ to cite this article : p. m. hajac et al .
, c. r. acad .
sci .
paris , ser .
i 336 ( 2003 ) .
_ 0.5 * rsum * 0.5*modules anti - yetter - drinfeld stables * nous dfinissons et tudions une classe de modules enlacs ( modules anti - yetter - drinfeld stables ) qui servent de coefficients pour lhomologie et la cohomologie hopf - cyclique . en particulier ,
nous expliquons leurs liens avec les modules de yetter - drinfeld et les doublets de drinfeld .
parmi les sources dexemples de modules anti - yetter - drinfeld stables , nous trouvons des extensions de hopf - galois munies dune version transpose de laction de miyashita - ulbrich .
_ pour citer cet article : p. m. hajac et al .
, c. r. acad .
sci .
paris , ser .
i 336 ( 2003 ) . |
the model system usually considered for the description of the stationary configurations of sandpiles on a tray table is the monge - kantorovich type system of pdes @xmath2 ( see , e.g. , @xcite ) .
the data of the problem are the flat surface of the table @xmath3 , the profile of the tray table @xmath4 , and the density of the source @xmath5 .
the dynamical behavior of the granular matter is pictured by the pair @xmath6 , where @xmath7 is the profile of the standing layer , whose slope has not to exceed a critical value ( @xmath8 ) in order to prevent avalanches , while @xmath9 is the thickness of the rolling layer ( see figure [ fig : sand ] ) .
the condition @xmath10 corresponds to require that the matter runs down only in the region where the slope of the heaps is maximal . on the border of the table
the standing layer fills the gap with the height @xmath4 , and the exceeding sand falls down . if @xmath11 , the model reduces to the open table problem , largely investigated in recent years ( see e.g. @xcite and the references therein ) . in the present paper
we consider a more general system of pdes in the open , bounded and connected set @xmath12 @xmath13 in the unknowns @xmath14 , @xmath15 .
the data are : * @xmath16 , a compact convex set of class @xmath17 containing the origin in its interior , and its ( strictly convex ) polar set @xmath18 .
the set @xmath19 appears in ( [ f : mkint ] ) by means of its gauge function @xmath20
. moreover the closure of @xmath1 will be equipped with a ( possibly asymmetric ) geodesic distance @xmath21 given in terms of the gauge function @xmath22 of the polar set @xmath18 ( see section [ s : extens ] ) ; * @xmath23 ; * @xmath24 , @xmath25-lipschitz function w.r.t .
@xmath21 .
this general version takes into account the possibility of homogeneous anisotropies , and can be applied , for example , to the bean s model for the description of the macroscopic electrodynamics of hard superconductors ( see @xcite , for the case @xmath11 and @xmath1 , @xmath19 smooth sets ) , and to optimal shape design ( see @xcite ) .
s. bianchini in @xcite proved that the lax
hopf function @xmath26 ( defined in ( [ f : lh ] ) below ) , coupled with an explicit @xmath27 solves ( [ f : mkint ] ) ( see theorem [ f : bian ] ) .
the function @xmath28 is constructed as follows .
the set @xmath1 can be covered by a family of disjoint transport rays , and every @xmath29-component of a solution to ( [ f : mkint ] ) satisfies a first order linear ode along almost every ray .
the function @xmath28 is then uniquely defined along each ray , by requiring that it vanishes at the final point .
this procedure gives a solution to ( [ f : mkint ] ) , thanks to a disintegration formula for the lebesgue measure along the rays . in this paper
we shall provide a complete characterization of the solutions @xmath6 to ( [ f : mkint ] ) .
starting from the fact that @xmath30 is a solution , we show that a pair @xmath6 is a solution if and only if both @xmath31 and @xmath32 are solutions , so that we can decouple the original problem ( see section [ s : minimum ] ) .
moreover , the admissible @xmath7-components are characterized as the solutions of a minimization problem which has an explicit minimal solution @xmath33 depending on the source term @xmath34 .
since @xmath26 is the maximal solution of the problem , the uniqueness of the @xmath7-component corresponds to the case when @xmath35 .
we prove that this can occur if and only if the final points in @xmath1 of the transport rays are contained in the support of the source @xmath34 ( see section [ s : uniqueu ] ) . concerning the @xmath29-component , in @xcite
it is proved that every admissible @xmath29 has to vanish at the final points in @xmath1 of the transport rays .
hence , by the very definition of @xmath28 , it is the unique @xmath29-component if the union @xmath36 of the rays with both endpoints on @xmath37 has null lebesgue measure .
we shall prove that the converse is also true : if the set @xmath36 has positive lebesgue measure , we can construct a nontrivial function @xmath14 such that @xmath38 , so that the uniqueness of the @xmath29-component fails .
these results generalize those in @xcite to the case when the reference set @xmath1 is possibly non regular , the boundary datum @xmath4 is possibly non - homogeneous , and the constraint @xmath19 is a possibly non - strictly convex set . in this case
we have to use a completely different strategy with respect to the previous approaches .
namely , if @xmath1 and @xmath19 are smooth enough , then it is possible to define a suitable notion of ( anisotropic ) curvatures of @xmath37 , and the analysis of solutions to ( [ f : mkint ] ) can be performed using the methods of differential geometry ( see , e.g. , @xcite ) . in the non - smooth case
these methods do not work , and a different approach is needed , based on the techniques developed in optimal transport theory ( see @xcite ) .
namely , system ( [ f : mkint ] ) can be understood as the pde formulation of an optimal transport problem with strictly convex cost function @xmath22 , assigned initial distribution @xmath39 and transport potential given by the lax - hopf function @xmath26 ( so that the final mass distribution @xmath40 is the measure concentrated on @xmath37 induced by the transportation of @xmath41 along the rays associated to @xmath26 ) . from this point of view
, the minimization problem considered in section [ s : minimum ] turns out to be the dual formulation of the monge problem ( see , e.g. , @xcite for an exhaustive presentation of this subject ) .
a comment on the regularity assumption on @xmath19 is in order .
namely , if @xmath19 is of class @xmath17 , then its polar set @xmath18 need not be of class @xmath17 , but it is strictly convex .
it is well - known that the strict convexity of @xmath18 , which is the unit ball of the dual norm @xmath22 , allows non - trivial simplifications in the proofs ; for example , the direction of the transport rays coincides with @xmath42 almost everywhere in @xmath1 .
related results in the case of a general possibly non - regular constraint @xmath19 can be found in @xcite .
the paper is organized as follows . in section [ s : prel ] we collect all the notation and preliminary results . in section [ s : extens ] we recall some generalization of the mcshane s extension theorem for lipschitz functions with constrained gradient . these general results will be used for introducing the maximal and the minimal extension of the boundary datum @xmath4 consistent with the gradient constraint .
section [ s : existence ] is devoted to the statement of the problem , some remarks on the hypotheses , and the statement of the existence result due to s. bianchini ( see ( * ? ? ?
one of the main tools needed for proving the uniqueness results is the fact that ( [ f : mkint ] ) is the euler - lagrange necessary condition for a minimization problem with a gradient constraint .
this property is proved in section [ s : minimum ] .
the necessary and sufficient conditions for the uniqueness of the solutions @xmath6 to ( [ f : mkint ] ) are proved in sections [ s : unique ] and [ s : uniqueu ] . as a byproduct of our analysis , with a little additional effort
, we provide a complete description of the singular sets considered in this kind of optimal transport problems
. * acknowledgements . *
the authors wish to thank the anonymous referee for valuable comments that have improved the presentation of the manuscript .
the standard scalar product of @xmath43 will be denoted by @xmath44 , while @xmath45 will denote the euclidean norm of @xmath46 . concerning the segment joining @xmath46 with @xmath47 , we set @xmath48\ } , \qquad \oray{x , y } : = \cray{x , y}\setminus\{x , y\}. \ ] ] given a set @xmath49 , its interior , its closure and its boundary will be denoted by @xmath50 , @xmath51 and @xmath52 respectively .
a bounded open set @xmath53 ( or , equivalently , @xmath54 or @xmath55 ) is of class @xmath56 , @xmath57 , if for every point @xmath58 there exists a ball @xmath59 and a one - to - one mapping @xmath60 such that @xmath61 , @xmath62 , @xmath63 , @xmath64 .
let us now fix the notation and the basic results concerning the convex set which plays the role of gradient constraint for the @xmath7-component in ( [ f : mkint ] ) .
in the following we shall assume that @xmath65 let us denote by @xmath18 the polar set of @xmath19 , that is @xmath66 we recall that , if @xmath19 satisfies ( [ f : hypk ] ) , then @xmath18 is a compact , strictly convex subset of @xmath67 containing the origin in its interior , and @xmath68 ( see , e.g. , @xcite ) .
the gauge function @xmath69 of @xmath19 is defined by @xmath70 it is straightforward to see that @xmath71 is a positively 1-homogeneous convex function such that @xmath72 .
the gauge function of the set @xmath18 will be denoted by @xmath22 .
the properties of the gauge functions needed in the paper are collected in the following theorem .
[ t : sch ] assume that @xmath16 satisfies ( [ f : hypk ] ) .
then the following hold : \(i ) @xmath71 is continuously differentiable in @xmath73 , and @xmath74
\(ii ) @xmath18 is strictly convex , and @xmath75 \(iii ) for every @xmath76 , @xmath77 belongs to @xmath78 , and @xmath79 see @xcite , section 1.7 . in what follows
we shall consider @xmath67 endowed with the possibly asymmetric norm @xmath80 , @xmath43 .
by theorem [ t : sch](ii ) , the unit ball @xmath18 of @xmath22 is strictly convex but , under the sole assumption ( [ f : hypk ] ) , it need not be differentiable .
moreover , the minkowski structure @xmath81 is not a metric space in the usual sense , since @xmath22 need not be symmetric ( for an introduction to nonsymmetric metrics see @xcite ) . finally , since @xmath18 is compact and @xmath82 , then the convex metric is equivalent to the euclidean one , that is there exist @xmath83 , @xmath84 such that @xmath85 for every @xmath86 .
the main motivation for introducing the convex metric associated to @xmath22 is the fact that the sobolev functions with the gradient constrained to belong to @xmath19 are the locally 1-lipschitz functions with respect to @xmath22 , as stated in the following result .
[ l : diseqrho ] let @xmath53 be a nonempty open set , and assume that the set @xmath87 satisfies ( [ f : hypk ] ) .
let @xmath22 be the gauge function of @xmath18 .
then the following properties are equivalent .
* @xmath88 is a locally @xmath25-lipschitz function with respect to @xmath22 , i.e. @xmath89 * @xmath90 , and @xmath91 for a.e .
@xmath92 .
let @xmath7 be a locally 1-lipschitz function in @xmath93 with respect to the metric @xmath22 .
then @xmath7 is locally lipschitz in @xmath93 , and if @xmath94 is a differentiability point of @xmath7 , by ( [ f : lipur ] ) we have @xmath95 hence @xmath96 for all @xmath97 , that is @xmath98 . since @xmath7 is a locally lipschitz function in @xmath93 with gradient in the bounded set @xmath19 , ( ii ) holds .
conversely , let @xmath7 satisfy ( ii ) , and let @xmath99 .
clearly , we can assume @xmath100 .
let @xmath101 , and let @xmath102 be the cylinder of radius @xmath103 around the segment @xmath104 .
let us choose @xmath105 such that @xmath106 .
by ( ii ) there exists a set @xmath107 with @xmath108-lebesgue measure zero , such that for every @xmath109 , @xmath7 is absolutely continuous along the segment @xmath110 , and @xmath111 for a.e .
@xmath112 $ ] .
let @xmath113 , @xmath114 .
then , for every @xmath115 , @xmath116 the conclusion now follows from the continuity of @xmath7 in @xmath93 , passing to the limit in @xmath117
. in the sequel we need to consider locally 1-lipschitz functions with respect to @xmath22 that agree with a given continuous boundary datum .
this can be done only if the boundary values are compatible with the gradient bounds , and , in absence of assumptions on the regularity of the reference set , the compatibility condition concerns the variations of the data along every admissible path .
here we introduce some notation for the curves , while the compatibility condition will be discussed in section [ s : extens ] .
[ d : gamma ] let @xmath118 be a nonempty compact set .
for every @xmath119 , @xmath120 will denote the ( possibly empty ) family of absolutely continuous curves @xmath121\to d$ ] such that @xmath122 and @xmath123 .
for every absolutely continuous curve @xmath121\to \r^n$ ] , let us denote by @xmath124 its length with respect to the convex metric associated to @xmath22 , that is @xmath125 if @xmath126 is a compact subset of @xmath67 , by a standard compactness argument we have that if @xmath127 , then there exists a geodesic in @xmath126 joining @xmath128 to @xmath129 , i.e. there exists a curve @xmath130 such that @xmath131 for every @xmath132 ( see e.g. ( * ? ? ? * thm .
4.3.2 ) , @xcite ) .
for sake of completeness we briefly recall that any geodesic curve lying in the interior of @xmath126 is in fact a segment , due to strict convexity of @xmath22 .
[ l : geoint ] let @xmath118 be a nonempty compact set .
let @xmath119 be two points such that @xmath133 , and let @xmath134 be a geodesic . if @xmath135 for every @xmath136 , then the support of @xmath137 is the segment @xmath138 .
by an approximation argument we can assume that @xmath139 so that @xmath140 for every @xmath141 $ ] .
let @xmath142 , @xmath141 $ ] , be a parametrization of the segment @xmath138 .
assume by contradiction that the support of @xmath137 is not the segment @xmath138 .
since @xmath18 is strictly convex we have that @xmath143 .
since the support of @xmath137 is contained in @xmath144 , there exists @xmath145 such that the support of @xmath146 is contained in @xmath144
. then @xmath147 \\ & = ( 1-\varepsilon)\len{\gamma } + \varepsilon\len{\eta } < \len{\gamma } , \end{split}\ ] ] in contradiction with the minimality of @xmath137 .
in this section , @xmath148 and @xmath126 will be two sets such that @xmath149 so that @xmath150 is an open ( possibly empty ) set . for a given continuous function @xmath151 , we want to discuss the existence of a locally lipschitz extension of @xmath152 in @xmath126 with the gradient constrained to belong to a convex set @xmath19 .
it is well known that , in the case @xmath153 , @xmath1 bounded open subset of @xmath67 , @xmath154 , and @xmath19 satisfying ( [ f : hypk ] ) , a necessary and sufficient condition for the existence of such an extension is @xmath155 where @xmath156 is the family of absolutely continuous curves @xmath121\to \overline{\omega}$ ] such that @xmath157 for every @xmath158 , @xmath159 and @xmath160 ( see ( * ? ? ?
* sect . 5 ) and @xcite ) .
moreover , if @xmath1 has a lipschitz boundary , one can relax the above necessary and sufficient condition by taking paths in @xmath161 ( see @xcite ) . in order to handle more general situations
, we will consider a continuous function @xmath151 satisfying @xmath162 where @xmath163 is the family of curves introduced in definition [ d : gamma ] .
this condition can be rephrased in the setting of the length space @xmath164 , where @xmath165 $ ] is the ( possibly asymmetric ) distance defined by @xmath166 ( we refer to ( * ? ? ?
2 ) for the basic properties of length spaces ; see in particular remark 2.2.6 for the extension to possibly asymmetric metrics . )
namely , ( [ f : h4 ] ) is equivalent to requiring that @xmath152 is a @xmath25-lipschitz function with respect to the @xmath21 . in the following lemma we recall the well known fact that ( [ f : h4 ] ) turns out to be a sufficient condition for the existence of minimal and maximal @xmath25-lipschitz extensions ( w.r.t .
@xmath21 ) of @xmath152 . notice that ( [ f : h4 ] ) is no more a necessary condition , as it can be seen when @xmath126 is a closed segment in @xmath167 .
[ l : genprop ] let @xmath148 , @xmath126 satisfy ( [ f : cd ] ) , let @xmath19 satisfy ( [ f : hypk ] ) , and let @xmath168 be a continuous function satisfying ( [ f : h4 ] ) . then the functions @xmath169 are continuous in @xmath126 , and @xmath170 in @xmath148 .
moreover , @xmath171 and @xmath172 a.e . in @xmath173 . as a consequence of lemma [ l : genprop ] ,
the functions @xmath174 and @xmath175 belong to the functional space @xmath176 it should be noted that in the definition of @xmath177 the infimum is taken over all paths in @xmath126 joining @xmath178 to @xmath46 , while the supremum in the definition of @xmath175 is taken over all paths in @xmath126 joining @xmath46 to @xmath178 .
this asymmetry is needed in order to construct exactly the maximal and the minimal elements of @xmath179 , as it is stated in lemma [ l : vischar ] below ( see ( * ? ? ?
2.1 ) ) .
[ d : visible ] given @xmath180 , we shall say that @xmath178 is a visible point for @xmath46 if the segment @xmath181 is entirely contained in the open set @xmath150 .
we shall denote by @xmath182 the set of all visible points for @xmath46 , that is @xmath183 .
[ l : vischar ] the functions @xmath174 , @xmath175 defined in ( [ f : psi ] ) are characterized in the following way : @xmath184 moreover @xmath185 for every @xmath186 .
[ c : penduno ] let @xmath187 be a point of differentiability for @xmath174 ( resp .
@xmath175 ) . then @xmath188 ( resp .
@xmath189 ) belongs to @xmath190 .
given @xmath187 , let @xmath191 be such that @xmath192 setting @xmath193 , the fact that @xmath194 and lemma [ l : diseqrho ] imply that @xmath195\,.\ ] ] then we have @xmath196 and hence @xmath197\,.\ ] ] if @xmath174 is differentiable at @xmath46 , ( [ f : linpsi ] ) implies that @xmath198 , so that @xmath199 on the other hand , by lemma [ l : genprop ] , @xmath200 , hence @xmath188 belongs to @xmath190 .
the proof for @xmath175 is omitted , since it is entirely similar .
our main goal is to provide a full analysis of the monge - kantorovich system of pdes @xmath201 under the following assumptions : * @xmath1 is a nonempty bounded connected open subset of @xmath67 , @xmath202 .
* @xmath203 , @xmath5 a.e . in @xmath1 .
* @xmath69 is the gauge function of a set @xmath87 satisfying ( [ f : hypk ] ) .
* @xmath204 is a continuous function , satisfying the condition @xmath205 where @xmath206 , \overline{\omega}):\ \gamma(0 ) = y,\ \gamma(1 ) = x \right\}\,.\ ] ] in the first equation in ( [ f : mk ] ) we understand @xmath207 defined also at the origin since , by the third condition , @xmath208 a.e . where @xmath209 .
the functional setting for the unknowns @xmath6 in ( [ f : mk ] ) is @xmath210 , where @xmath211 and @xmath212 by a solution of ( [ f : mk ] ) we mean a pair @xmath213 satisfying @xmath214 and @xmath215 a.e . in @xmath1 .
let us define the lax - hopf function @xmath216 by @xmath217 which corresponds to the function @xmath174 defined in ( [ f : lh3 ] ) in the special case @xmath153 , @xmath218 , and @xmath219 ( observe that @xmath220 and @xmath221 ) . as a consequence of lemmas [ l
: genprop ] , [ l : vischar ] , and corollary [ c : penduno ] , we obtain the following result , which is well - known when @xmath1 is a lipschitz domain ( see e.g. ( * ? ? ?
* chap . 5 ) ) .
[ t : genprop ] if ( h1 ) , ( h3 ) and ( h4 ) are fulfilled , then @xmath222 and @xmath223 a.e . in @xmath1 .
moreover , @xmath26 is the maximal element in @xmath224 , and it is characterized by @xmath225 where @xmath226 . if @xmath11 , the lax
hopf function @xmath227 is the distance from @xmath228 with respect to the convex metric associated to @xmath22 , that is @xmath229 , @xmath230 ( see @xcite for a complete treatment of this subject ) .
more generally , if @xmath231 satisfies @xmath232 the lax
hopf function @xmath26 coincides with @xmath233 namely , from ( [ f : onelip ] ) it is easy to check that @xmath234 on @xmath228 .
moreover , from the representation formula ( [ f : lh2 ] ) it is clear that @xmath235 . on the other hand ,
let @xmath236 and let @xmath237 be such that @xmath238 .
let @xmath239 be such that @xmath240 . then , by the representation formula ( [ f : lh2 ] ) and by ( [ f : onelip ] ) we have that @xmath241 so that @xmath242 .
the following example shows that if we merely require ( h4 ) , then in general @xmath243 .
given @xmath244 let us define the set @xmath245 and let @xmath231 be defined by @xmath246 .
if we choose @xmath247 , then @xmath248 for every @xmath249 , and @xmath124 is the usual ( euclidean ) length of a curve @xmath137 .
a straightforward computation shows that @xmath4 satisfies ( h4 ) . on the other hand
, it is also clear that @xmath4 does not satisfy ( [ f : onelip ] ) .
namely , if we consider the two boundary points @xmath250 , @xmath251 , then we have @xmath252 in this case , the function @xmath253 defined in ( [ f : lh4 ] ) does not belong to @xmath224 , since @xmath254 and hence it does not achieve the boundary value at @xmath255 .
we conclude this section with the following result due to s. bianchini ( see @xcite and @xcite for its generalization to the case of nonsmooth constraint sets @xmath19 ) .
[ t : exv ] there exists a weak solution @xmath27 to the transport equation @xmath256 as a corollary of theorems [ t : genprop ] and [ t : exv ] we conclude that @xmath30 is a solution to ( [ f : mk ] ) .
in this section we shall discuss the connection between the system of pdes ( [ f : mk ] ) and the variational problem @xmath257 it is clear that , since @xmath34 is non - negative and @xmath26 is the maximal element in @xmath224 , then @xmath26 is a solution to ( [ f : min ] ) . we shall show that ( [ f : mk ] ) is the euler - lagrange necessary and sufficient condition for the minimum problem ( [ f : min ] ) .
this property , which is the analogous to the duality in mass transport theory , will be useful in the proofs of the uniqueness of the solutions of ( [ f : mk ] ) . as a preliminary step ,
we need to enrich the class of test functions allowed in ( [ f : weak ] ) .
[ l : approx ] let @xmath258 , @xmath259 on @xmath228 .
then there exists a sequence @xmath260 such that @xmath261 and @xmath262 a.e . in @xmath1 , and @xmath263 , @xmath264 for every @xmath115 .
we use a standard truncation argument .
let @xmath265 such that @xmath266 for every @xmath267 , @xmath268 for @xmath269 , @xmath270 for @xmath271 , @xmath272 for every @xmath267 .
let us define @xmath273 , so that @xmath274 pointwise in @xmath275 .
moreover , @xmath276 , hence @xmath277 a.e . in @xmath1 and @xmath278 .
let @xmath279 be the standard family of mollifiers in @xmath67 .
since @xmath280 has compact support in @xmath1 , we can choose a sequence @xmath281 , such that the sequence @xmath282 has the required properties .
[ c : density ] @xmath283 is dense in the space @xmath284 with respect to the weak@xmath285 topology of @xmath286 . as a consequence ,
if @xmath213 satisfies ( [ f : weak ] ) , then @xmath287 we are now in a position to prove that the monge - kantorovich system ( [ f : mk ] ) is the euler - lagrange condition for the minimum problem ( [ f : min ] ) . [
t : equiv ] the minimum problem ( [ f : min ] ) and the system of pdes ( [ f : mk ] ) are equivalent in the following sense .
* @xmath288 is a solution to ( [ f : min ] ) if and only if there exists @xmath289 such that @xmath6 is a solution to ( [ f : mk ] ) .
* let @xmath288 be a solution to ( [ f : min ] ) .
then @xmath6 is a solution to ( [ f : mk ] ) if and only if @xmath32 is a solution to ( [ f : mk ] ) .
let us denote by @xmath290 the indicator function of the set @xmath19 , that is @xmath291 so that @xmath292\ , dx = -\int_{\omega}fu\ , dx , \qquad \forall u\in \spacex\,.\ ] ] since @xmath71 is differentiable in @xmath293 , the subgradient of the indicator function can be explicitly computed , obtaining @xmath294 ( see e.g. ( * ? ? ?
23 ) ) . hence ,
if @xmath6 is a solution to ( [ f : mk ] ) , then @xmath295 for a.e .
@xmath236 , so that , for every @xmath296 @xmath297 where the last equality follows from corollary [ c : density ] .
this proves that @xmath7 is a solution to ( [ f : min ] ) .
assume now that @xmath288 is a minimizer for @xmath298 , so that @xmath299 a.e . in @xmath1 ,
due to the maximality of @xmath26 in @xmath224 and the fact that @xmath5 . by corollary
[ c : density ] we can choose @xmath300 as test function in the transport equation ( [ f : dive ] ) solved by @xmath26 , getting @xmath301 on the other hand , by theorem [ t : sch](ii ) and the fact that @xmath302 a.e . in @xmath1 , we have @xmath303 so that @xmath304 and @xmath305 a.e . in @xmath1 , that is @xmath31 is a solution of ( [ f : mk ] ) .
this concludes the proof of ( i ) .
let us prove ( ii ) .
the previous computation shows that if @xmath306 is a solution of ( [ f : mk ] ) , and @xmath288 is a solution of the minimum problem ( [ f : min ] ) , then also @xmath6 is a solution of ( [ f : mk ] ) .
finally , let @xmath6 be a solution to ( [ f : mk ] ) . upon observing that @xmath307 a.e . in @xmath1 , and choosing @xmath300 as test function in the weak formulation of the transport equation @xmath308
we conclude that the opposite implication holds .
thanks to theorem [ t : equiv](ii ) , the uniqueness of the @xmath29 component of the solution of the system of pdes ( [ f : mk ] ) will follow from the fact that the function @xmath309 introduced in theorem [ t : exv ] is the unique weak solution in @xmath310 of the equation @xmath311 we shall use an explicit representation of @xmath309 proved in @xcite and , in order to explain it , we need to introduce some additional tools , related to the directions where the function @xmath26 has the maximal slope . recalling the representation formula ( [ f : lh2 ] ) for @xmath26 , for every @xmath312 let us define the projections of @xmath46 by @xmath313 and , for every @xmath236 , let @xmath314 be the set of directions through @xmath46 @xmath315
let @xmath316 be the set of points with multiple projections , that is @xmath317 and for every @xmath318 , let @xmath319 and @xmath320 denote the unique elements in @xmath321 and @xmath314 respectively .
finally , let @xmath322 be defined by @xmath323\ } , & x\in \omega\setminus d\ , , \\ 0 & x\in d\ , , \end{cases}\ ] ] and let @xmath324 be the set @xmath325 where we understand that @xmath326 if @xmath180 .
we shall call _ transport ray _ through @xmath236 any segment @xmath327 , @xmath328 .
it is clear from the definition that , if @xmath318 , then there is a unique transport ray @xmath329 through @xmath46 . on the other hand , if @xmath180 , any segment @xmath330 , with @xmath328 , is a transport ray through @xmath46 .
the transport rays correspond to the segments where @xmath26 grows linearly with maximal slope , as stated in the following lemma .
hence in the sandpiles model , they correspond to the directions where the matter is allowed to run down , that is where the @xmath29component could be nonzero .
[ l : prunica ] for every @xmath236 and @xmath331 we have @xmath332 moreover , if @xmath333 , then @xmath334 the identity ( [ f : ulin ] ) follows from ( [ f : linpsi ] ) , while ( [ f : ulin2 ] ) for @xmath333 follows from ( [ f : ulin ] ) when @xmath335 , and from the very definition of @xmath336 when @xmath337 .
[ l : project ] given @xmath318 , let @xmath329 be the transport ray through @xmath46 .
then @xmath338 for every @xmath339 . from ( [ f : ulin2 ] ) it is clear that @xmath340 for every @xmath341 .
assume by contradiction that there exists @xmath339 such that @xmath342 contains a point @xmath343 .
we claim that the three points @xmath47 , @xmath344 and @xmath336 can not be aligned .
namely , we can not have @xmath345 , for otherwise we would have @xmath346 . on the other hand
, we can not have @xmath347 : indeed , in this case , by ( [ f : ulin2 ] ) we would have @xmath348 , while , by ( [ f : ulin ] ) , we know that @xmath349 . putting together these information , we get @xmath350 a contradiction . since the curve @xmath351 , joining @xmath47 to @xmath336 and lying in @xmath1 except for its endpoints , is not a segment , by lemma [ l : geoint ] it can not be a geodesic , so that @xmath352 taking into account ( [ f : ulin2 ] ) we get @xmath353 a contradiction with @xmath354 . at this stage
we are dealing with three meaningful sets of singular points associated to @xmath26 : * the sets of points in @xmath1 where @xmath26 is not differentiable ; * the sets of points in @xmath1 with multiple projections ; * the set of endpoints in @xmath1 of the transport rays .
the set @xmath324 plays a central role in optimal transport theory . as a matter of fact ,
the disintegration of the lebesgue measure in the reference set @xmath1 along transport rays , can be achieved only if @xmath324 has lebesgue measure zero , and this is one of the main tools of the theory .
if @xmath1 and @xmath19 are sufficiently smooth ( e.g. both of class @xmath355 , and @xmath19 with strictly positive curvatures ) , one can prove that @xmath126 coincides with @xmath356 ( see , e.g. @xcite ) , so that , by rademacher s theorem , @xmath126 has zero lebesgue measure .
moreover , using the methods of differential geometry , it is possible to prove that @xmath357 and to characterize @xmath358 in terms of optimal focal points . as a consequence one can prove that @xmath324 has zero lebesgue measure and give detailed rectifiability results ( see , e.g. , @xcite )
. if @xmath1 and @xmath19 are not regular , the set @xmath359 may have positive lebesgue measure ( see @xcite for an example where @xmath360 , @xmath361 , and @xmath19 is the unit ball ) , and @xmath324 can be a proper subset of @xmath359 .
we recall here some known properties of the singular sets , adding some remarks in order to complete the description of the relationships between these sets .
the known results , proved in @xcite , are collected in the following proposition .
[ p : d ] the sets @xmath356 , @xmath126 and @xmath324 satisfy the following properties .
* @xmath362 .
more precisely , if @xmath26 is differentiable at @xmath236 , then @xmath46 has a unique projection and @xmath363 .
* the set @xmath324 has zero lebesgue measure .
* @xmath126 is a @xmath108 rectifiable set .
[ r : exa ] we recall that @xmath364 if @xmath19 is strictly convex . proposition [ p : d](i ) states that @xmath362 , in general .
the following example shows that it can happen that @xmath365 .
let us consider the set @xmath366 and let @xmath19 be its polar set .
the set @xmath18 is strictly convex but it is not of class @xmath17 , hence the set @xmath19 is of class @xmath17 but it is not strictly convex ( see theorem [ t : sch ] ) .
let @xmath1 be the hexagon @xmath367 ( see figure [ fig : exa ] ) . the lax
hopf function @xmath227 associated to @xmath19 and to the boundary datum @xmath11 is the distance function to the boundary of @xmath1 with respect to the convex metric @xmath22 , which can be explicitly computed as @xmath368 so that @xmath369 . and
@xmath18 of example [ r : exa],height=151 ] since @xmath126 is a subset of @xmath324 , then by proposition [ p : d](ii ) it has lebesgue measure zero .
notice that @xmath126 can be a proper subset of @xmath324 even in the regular case .
for example , let @xmath1 be the ellipsis @xmath370 , where @xmath371 , and let @xmath372 .
the points @xmath373 and @xmath374 are the centers of curvature of @xmath228 at @xmath375 and @xmath376 respectively
. it can be checked that @xmath377 whereas @xmath378 . moreover , since @xmath324 need not be closed ( see ( * ? ? ?
* remark 5.2 ) ) , @xmath379 in general .
the following proposition shows that @xmath380 .
[ p : bu ] for every @xmath381 we have that @xmath382 , where @xmath383 is the function defined in ( [ f : tau ] ) .
let us fix @xmath384 .
we can assume without loss of generality that @xmath385 , and @xmath386 , @xmath387 .
since the map @xmath388 is continuous in the open set @xmath389 ( see ( * ? ? ?
3.2 ) ) , there exists @xmath105 such that @xmath390 , and @xmath391 for every @xmath392 .
given @xmath393 , let us define the function @xmath394 by @xmath395 so that @xmath396 . hence @xmath152 maps @xmath397 into the hyperplane @xmath398 .
the continuity of @xmath399 in @xmath397 ensures that we can choose @xmath400 in such a way that @xmath401 for every @xmath402 . summarizing
, @xmath403 is a continuous map , @xmath404 , and @xmath405 belongs to the ray @xmath406 .
moreover , for every @xmath407 , we have that @xmath408 , hence @xmath409 and , by corollary [ l : project ] , we infer that @xmath410 . since @xmath152 maps @xmath411 into the @xmath108-dimensional hyperplane @xmath412 , by the borsuk
ulam lemma ( see ( * ? ? ?
* cor . 4.2 ) ) there exists @xmath413 such that @xmath414 .
furthermore @xmath415 , so that the equality @xmath414 implies that @xmath416 i.e. , @xmath417 is parallel to @xmath418 .
but the transport ray through @xmath417 and @xmath419 contains the origin , hence @xmath420 and @xmath421 . for the reader
s convenience we collect here the results concerning the relationships between the singular sets .
the sets @xmath356 , @xmath126 and @xmath324 have zero lebesgue measure .
moreover , @xmath422 and @xmath423 , with possibly strict inclusions .
we are now in a position to briefly recall the explicit representation of the function @xmath309 appearing in ( [ f : dive ] ) , one of the main results in @xcite .
[ f : bian ] there exists a positive function @xmath424 , absolutely continuous along almost every transport ray , such that an explicit solution @xmath27 of ( [ f : dive ] ) is given by @xmath425 more precisely , for a.e .
@xmath333 the function @xmath28 is locally absolutely continuous along the ray @xmath426 , @xmath427 , and satisfies @xmath428 = - f(x+t d(x))\ , \alpha(x+t d(x)),\\ \displaystyle \lim_{t\to \tau(x)^- } \vf(x+t d(x))\ , \alpha(x+t d(x ) ) = 0 .
\end{cases}\ ] ] see @xcite , theorems 5.7 and 7.1 .
the reader may want to compare the representation formula ( [ f : vf ] ) with the one proved in the regular homogeneous case ( see @xcite ) , where @xmath429 can be explicitly written in terms of the distance from the boundary @xmath430 and of the anisotropic curvatures @xmath431 of @xmath228 at @xmath319 .
more precisely , one has @xmath432 , \qquad x\in\omega\setminus d\,.\ ] ] we shall find a necessary and sufficient condition for having that @xmath28 is the unique solution to ( [ f : divei ] ) .
the following example shows why the uniqueness of the @xmath29-component may fail .
[ e : square ] let @xmath433 , @xmath372 , and let @xmath434 on @xmath37 , so that @xmath435 in @xmath1 .
then every function of the form @xmath436 , @xmath437 for @xmath438 $ ] , is a non - negative solution of the equation @xmath439 .
notice that in this case @xmath440 in @xmath1 , so that ( [ f : divei ] ) has infinitely many solutions of the form @xmath441 , with @xmath442
. the existence of non - trivial variations of @xmath28 is due to the fact that @xmath1 is covered by transport rays with both endpoints on @xmath228
. now it should be clear that , in order to discuss the uniqueness of the @xmath29component of the solutions to ( [ f : divei ] ) , we need to introduce the set @xmath443 where the set @xmath444 is defined by @xmath445 in example [ e : square ] , @xmath446 and @xmath447 .
in general @xmath36 is the union of the transport rays ending on @xmath228 . on this
set every function in @xmath224 is forced to have maximal slope and hence to coincide with @xmath26 .
[ l : sinray ] let @xmath448 be given . then the following hold .
* @xmath449 and @xmath450 for every @xmath451 . * if @xmath452 , then @xmath453 for every @xmath454 . as a consequence every @xmath288 coincides with @xmath26 on the set @xmath36 .
let us start by proving that @xmath328 for every @xmath451 .
namely @xmath455 where the last inequality follows from lemma [ l : diseqrho ] . hence @xmath456
, so that @xmath328 .
actually we can prove that @xmath457 for every @xmath451 using the same argument of the proof of corollary [ l : project ] : if there exists @xmath458 , @xmath459 , then the three points @xmath460 , @xmath46 and @xmath461 can not be aligned , so that @xmath462 and @xmath463 a contradiction .
since @xmath464 and , for every @xmath465 , @xmath466 , it is clear that @xmath467 . finally , let @xmath468 and @xmath454 be given . then , using lemma [ l : diseqrho ] we get @xmath469 and ( ii ) follows .
[ t : uniqv ] the function @xmath28 defined in ( [ f : vf ] ) is the unique solution of ( [ f : divei ] ) if and only if the set @xmath36 has zero lebesgue measure .
the uniqueness of @xmath28 when @xmath36 has zero lebesgue measure is proved in ( * ? ? ?
the proof is based on the fact that every weak solution @xmath14 to ( [ f : divei ] ) is locally absolutely continuous along almost every transport ray , and satisfies @xmath470 = - f(x+t d(x))\ , \alpha(x+t d(x))\,.\ ] ] moreover , if the endpoint @xmath336 of the transport ray through @xmath46 belongs to @xmath1 , the @xmath471 satisfies the initial condition @xmath472 so that , by theorem [ f : bian ] , @xmath473 along that ray .
let us now assume that @xmath36 has positive lebesgue measure .
it is clear that the function @xmath474 is the lax - hopf function corresponding to the geometry induced by the convex set @xmath475 and the boundary datum @xmath476 .
assume that @xmath26 and @xmath477 are differentiable at a point @xmath465 , with @xmath448 .
from proposition [ p : d ] we have that @xmath478 .
the same argument , used with @xmath476 as boundary datum and the geometry induced by @xmath475 , shows that @xmath479 where @xmath480 and @xmath481 denote respectively the gauge functions of the convex sets @xmath475 and @xmath482 .
let us consider the indicator function of the set @xmath36 , @xmath483 by theorem [ t : exv ] , there exist two nonnegative functions @xmath484 , @xmath485 such that @xmath486 looking at the explicit representations of @xmath484 and @xmath485 as in ( [ f : vf ] ) , we easily get @xmath487 so that , for every @xmath488 , we obtain @xmath489 hence for every @xmath490 the function @xmath491 is a non - negative weak solution to ( [ f : divei ] )
. it can be easily checked that if the datum @xmath4 satisfies ( h4 ) with strict inequality holding for every @xmath492 , then the lebesgue measure of @xmath36 is zero , so that @xmath28 is the unique @xmath29-component allowed in ( [ f : mk ] ) .
if the lebesgue measure of @xmath493 is zero , then by lemma [
l : sinray ] we have that @xmath494 .
on the other hand , it can be easily checked that the converse implication is also true . in @xcite
( see also @xcite ) it is proved that either @xmath495 or there exists a weak solution @xmath496 to @xmath497 such that @xmath498 a.e . in @xmath1 . in the proof of theorem [ t : uniqv ]
we have shown that , as soon as @xmath36 has positive lebesgue measure , it is possible to construct a solution @xmath14 to ( [ f : homo ] ) such that @xmath499 a.e .
on @xmath36 .
in theorem [ t : equiv ] we have shown that the set of admissible @xmath7-components of the solutions to ( [ f : mk ] ) coincides with the set of solutions to the minimum problem ( [ f : min ] ) . in this section
we shall construct , using lemma [ l : genprop ] , the minimal solution to ( [ f : min ] ) , and we shall give a necessary and sufficient condition in order to have that this function equals the maximal solution @xmath26 in @xmath1 .
the function @xmath33 is the function @xmath175 defined in ( [ f : psi ] ) corresponding to @xmath153 , @xmath513 , and @xmath514 on @xmath148 .
clearly the function @xmath152 satisfies ( [ f : h4 ] ) on @xmath148 , since @xmath26 satisfies that condition on @xmath275 .
then the stated properties follow from lemmas [ l : genprop ] , [ l : vischar ] and [ l : sinray ] .
a function @xmath288 is a solution to ( [ f : min ] ) if and only if @xmath515 .
moreover , the function @xmath33 coincides with @xmath26 in @xmath1 ( and hence @xmath26 is the unique solution to ( [ f : min ] ) ) if and only if @xmath516 .
the first assertion follows from proposition [ p : uf](iii ) and from the fact that @xmath288 is a solution to ( [ f : min ] ) if and only if @xmath511 on @xmath502 .
let us prove the uniqueness result .
assume that @xmath516 , so that @xmath517 and , from proposition [ p : uf](ii ) , @xmath518 on @xmath324 .
let @xmath519 be given , and let @xmath520 be the endpoint of the ray through @xmath46 .
we have @xmath521 and hence @xmath509 .
assume now by contradiction that @xmath35 in @xmath1 and that there exists @xmath522 such that @xmath523 .
let @xmath524 and @xmath525 be such that @xmath526 notice that @xmath527 , otherwise we get @xmath528 . hence @xmath94 , @xmath47 and @xmath344 are three distinct points .
moreover we have that @xmath529 , otherwise by lemma [ l : geoint ] we should have @xmath530 and @xmath531 a contradiction .
cannarsa , p. cardaliaguet , g. crasta , and e. giorgieri , _ a boundary value problem for a pde model in mass transfer theory : representation of solutions and applications _ , calc .
partial differential equations * 24 * ( 2005 ) , 431457 .
cannarsa and c. sinestrari , semiconcave functions , hamilton - jacobi equations and optimal control , progress in nonlinear differential equations and their applications , vol .
58 , birkhuser , boston , 2004 .
gromov , metric structures for riemannian and non - riemannian spaces , english ed .
, modern birkhuser classics , birkhuser boston inc . ,
boston , ma , 2007 , based on the 1981 french original , with appendices by m. katz , p. pansu and s. semmes , translated from the french by sean michael bates .
mr mr2307192 ( 2007k:53049 ) li and l. nirenberg , _ the distance function to the boundary , finsler geometry and the singular set of viscosity solutions of some hamilton jacobi equations _ , commun .
pure appl . math .
* 58 * ( 2005 ) , 85146 .
villani , optimal transport , grundlehren der mathematischen wissenschaften [ fundamental principles of mathematical sciences ] , vol .
338 , springer - verlag , berlin , 2009 , old and new .
mr 2459454 ( 2010f:49001 ) | we prove a uniqueness result of solutions for a system of pdes of monge - kantorovich type arising in problems of mass transfer theory .
the results are obtained under very mild regularity assumptions both on the reference set @xmath0 , and on the ( possibly asymmetric ) norm defined in @xmath1 . in the special case
when @xmath1 is endowed with the euclidean metric , our results provide a complete description of the stationary solutions to the tray table problem in granular matter theory . |
various models for nuclear structure have been developed in order to study low energy phenomena of the atomic nuclei .
whereas straightforward application of the bare @xmath0 interaction is yet limited only to light nuclei @xcite , the nuclear structure seems to be well described by relatively simple effective interactions at low energies .
although the effective interactions may depend on the models , there should be basic characters in the effective interactions for the low energy phenomena , irrespective to the models .
on the other hand , since the invention of the secondary beam technology , experimental data on the unstable nuclei have disclosed new aspects of the nuclear structure .
a remarkable example is the dependence of magic numbers on the neutron excess @xcite . in regard to the new magic numbers discovered near the neutron drip line
, a question has been raised on a character of the effective interactions relating to the spin - isospin flip mode @xcite .
mean - field theories have successfully been applied to the nuclear structure problems , in particular for stable nuclei .
they are also useful to investigate basic characters of the effective interactions .
however , not many effective interactions have been explored for the nuclear mean - field calculations so far .
the skyrme interaction @xcite has been popular in the hartree - fock ( hf ) calculations , since the zero - range form is easy to be handled . among a limited number of finite - range interactions ,
the gogny interaction @xcite is widely applied to the mean - field calculations , in which the gaussian form is assumed for the central force .
the parameter - sets both of the skyrme and the gogny interactions have been adjusted mainly to the data on the nuclei around the @xmath1-stability . it is not obvious whether the available parameter - sets of these interactions account for the new magic numbers properly . in order to exploit effective interactions applicable also to unstable nuclei , guide from microscopic theories will be important .
brueckner s @xmath2-matrix has been a significant clue to studies in this course .
although microscopic approaches using the @xmath2-matrix have not yet been successful in reproducing the saturation properties , notable progress has been made recently . in the shell
model approaches , microscopic effective interactions have been shown to reproduce observed levels remarkably well @xcite .
it should be noted , however , that the shell model interactions are usually specific to mass regions , and their global characters have not been discussed in detail , despite several exceptions @xcite .
the so - called michigan 3-range yukawa ( m3y ) interaction @xcite has been derived from the bare @xmath0 interaction , by fitting the yukawa functions to the @xmath2-matrix . represented by the sum of the yukawa functions , the m3y - type interactions will be tractable in various models .
it has been shown that the m3y interaction gives matrix elements similar to reliable shell model interactions @xcite . moreover , with a certain modification , m3y - type interactions have successfully been applied to nuclear reactions @xcite . by using a recently developed algorithm @xcite ,
a class of the m3y - type interactions can be applied also to the mean - field calculations . under such circumstances
, it will be of interest to explore m3y - type interactions and to investigate their characters in the mean - field framework . in this article
, we shall develop m3y - type interactions and investigate their characters via the hf calculations .
nuclear effective hamiltonian consists of the kinetic energy and the effective interaction , @xmath3 here @xmath4 and @xmath5 are the indices of individual nucleons
. it will be natural to assume the effective interaction @xmath6 to be translationally invariant , except for the density dependence mentioned below .
we consider the effective interaction having the following form , @xmath7^\alpha \delta(\mathbf{r}_{12})\ , .
\label{eq : effint}\end{aligned}\ ] ] the relative coordinate is denoted by @xmath8 and @xmath9 .
correspondingly , the relative momentum is defined by @xmath10 .
@xmath11 is the relative orbital angular momentum , @xmath12 @xmath13 , @xmath14 are the nucleon spin operators , and @xmath15 is the tensor operator , @xmath16\ , .
\label{eq : tensor}\ ] ] @xmath17 represents an appropriate function of @xmath18 , the subscript @xmath19 corresponds to the parameter attached to the function ( _ e.g. _ the range of the interaction ) , and @xmath20 is the coefficient .
examples of @xmath17 are the delta , the gauss and the yukawa functions . @xmath21 ( @xmath22 ) denotes the spin ( isospin ) exchange operator , while @xmath23 , @xmath24 , @xmath25 and @xmath26 are the projection operators on the singlet - even ( se ) , triplet - even ( te ) , singlet - odd ( so ) and triplet - odd ( to ) two - particle states , respectively , which are defined by @xmath27 the nucleon density is denoted by @xmath28 .
the original m3y interaction is represented in the form of eq .
( [ eq : effint ] ) , with @xmath29 and @xmath30 . as discussed in ref .
@xcite , the skyrme and the gogny interactions are obtained by setting @xmath17 appropriately , except for some parameter - sets of the skyrme interaction in which certain terms are expressed only in the density - functional form .
the saturation of density and energy is a basic property of nuclei . in developing effective interactions adaptable for many nuclei ,
it is required to reproduce the saturation property .
however , the non - relativistic @xmath2-matrix fails to reproduce the saturation at the right density and energy .
therefore , it will not be appropriate to use the @xmath2-matrix for hf calculations without any modification , although several hf approaches using interactions derived from the @xmath2-matrix were tried in earlier studies @xcite .
the m3y interaction was obtained so that the @xmath2-matrix at a certain density could be reproduced by a sum of the yukawa functions .
the m3y interaction gives no saturation point within the hf theory , unless density - dependence is taken into account explicitly .
_ applied the m3y interaction to nuclear reactions in the folding model , by making the coupling constants dependent on densities @xcite . in their approach
the exchange terms are treated approximately .
however , the exchange terms may contribute quite significantly to the nuclear structure .
we here keep the coupling constants in @xmath31 independent of density , while introduce a density - dependent contact interaction ( @xmath32 in eq .
( [ eq : effint ] ) ) , as in the skyrme and the gogny interactions .
we can then treat the exchange ( _ i.e. _ the fock ) terms exactly with the currently available computers .
it should be mentioned that there has been an interesting attempt to approximate the exchange terms of the interaction in the density - matrix expansion @xcite , although the accuracy of the density - matrix expansion should be checked carefully .
we start from the paris - potential version of the m3y interaction @xcite .
this original parameter - set with no density - dependence is hereafter called ` m3y - p0 ' .
we shall modify this interaction so as to reproduce the saturation properties . in the isotropic uniform nuclear matter ,
matrix elements of @xmath33 and @xmath34 between the hf states vanish .
therefore @xmath35 determines the bulk properties such as the saturation .
the range parameters for the yukawa functions @xmath36 in @xmath31 are @xmath37 , @xmath38 and @xmath39 in the m3y interaction , which correspond to the compton wave - lengths of mesons with masses of about 790 , 490 and 140mev , respectively .
we do not change these parameters . for the longest - range part ( @xmath40 ) , the coupling constants @xmath41 , @xmath42 , @xmath43 and @xmath44 are fixed to be those of the one - pion - exchange potential ( opep ) , as in m3y - p0 .
the interaction @xmath32 in eq .
( [ eq : effint ] ) acts only on the se and te channels , @xmath45 microscopic investigations have shown that the density - dependence of the te part is primarily responsible for the saturation @xcite , as a higher - order effect of the tensor force . while the interaction in the se channel is attractive at low densities , it also has certain density - dependence originating in the strong short - range repulsion .
thus , a possible way of modifying the m3y interaction may be to replace a fraction of the repulsion in the se and te channels by @xmath32 .
in addition to the saturation properties which are relevant to the central force , the ls splitting is significant in describing the shell structure of nuclei . while true origin of the ls splitting is not yet obvious @xcite , ls splittings obtained from hf calculations with the @xmath2-matrix interaction are too small , in comparison with the observed ones . from the hf calculations for finite nuclei , we find that @xmath33 should be about twice as strong as that of m3y - p0 to reproduce the observed ls splittings .
the tensor force influences the ordering of the single - particle ( s.p . )
orbits . to reproduce the observed ordering , @xmath34 should be smaller than that of m3y - p0 .
we here introduce an overall enhancement factor to @xmath46 and an overall reduction factor to @xmath47 , as will be shown in section [ sec : dmprop ] . in this paper
we shall use two parameter - sets for modified m3y interaction , ` m3y - p1 ' and ` m3y - p2 ' , in order to show sensitivity to the parameters for some results . in m3y - p1
, we replace the shortest - range ( @xmath48 ) repulsive part of @xmath49 by @xmath32 in a simple manner .
we reduce both @xmath50 and @xmath51 by a single factor , keeping the se / te ratio in @xmath32 to be equal to @xmath52 in m3y - p0 , by imposing @xmath53 the reduction factor and @xmath54 are determined so as for the saturation density and energy in the nuclear matter to be typical values , as presented in the subsequent section .
characters of m3y - p1 will be investigated in the nuclear matter .
although this modification is too simple to reproduce properties of finite nuclei , the m3y - p1 set will be useful to clarify what characters arise from the original m3y interaction , relatively insensitive to the phenomenological modification . in the m3y - p2 set ,
all @xmath20 parameters belonging to the @xmath48 and @xmath55 channels in @xmath31 are shifted from those of m3y - p0 .
although we have three ranges in @xmath31 , the number of adjustable parameters is no greater than in the gogny interaction , since we fix the opep part .
we fit those parameters , together with the enhancement factor for @xmath33 and the reduction factor for @xmath34 , to the binding energies of several doubly magic nuclei .
the resultant values of the parameters will be shown later .
basic characters of nuclear effective interactions can be discussed via properties of the infinite nuclear matter ; in particular , properties at and around the saturation point . in this section
we investigate characters of the m3y - type interactions via the nuclear matter properties within the hf theory . in comparison
, we also discuss those of the skyrme and the gogny interactions .
we use the d1s parameter - set @xcite for the gogny interaction . in most of the skyrme
hf approaches , the ls currents arising from the momentum - dependence of the central force are ignored , and the parameters are adjusted without their contribution .
although this treatment occasionally improves some characters of the interactions , in this paper we would focus on characters of the two - body interactions , rather than those of density functionals . for this reason we adopt the sly5 set @xcite , which is devised for calculations including the ls currents . in the hf theory of the nuclear matter , the s.p .
wave - functions can be taken to be the plane wave , @xmath56 here @xmath57 ( @xmath58 ) denotes the spin ( isospin ) wave - function , and @xmath59 indicates the volume of the system , for which we will take the @xmath60 limit afterward .
energy for this state is defined as @xmath61 energy of the nuclear matter is expressed by a function of densities depending on the spin and the isospin , @xmath62 ( @xmath63 ; @xmath64 ) .
the density variables can be converted to the total density @xmath65 , and the spin- and isospin - asymmetry parameters @xmath66 where @xmath67 ( @xmath68 ) in the summation takes @xmath69 , corresponding to @xmath64 ( @xmath63 ) . by assuming that the s.p .
states are occupied up to the fermi momentum , the density is related to the fermi momentum for each spin and isospin , @xmath70 the total energy of nuclear matter is given by @xmath71 as already pointed out , only @xmath35 contributes to the energy of the isotropic nuclear matter . in appendix
[ app : nme ] , several formulae on the hf energy of the nuclear matter are derived for interactions expressed in the form of eq .
( [ eq : effint ] ) , with general and typical @xmath72 . as well as for the m3y - type interactions ,
the nuclear matter energies are calculated for the skyrme and the gogny interactions by using these formulae . in the spin - saturated symmetric nuclear matter , we have @xmath73 , which indicates @xmath74 and @xmath75 . in this case
we denote the fermi momentum simply by @xmath76 .
the lowest energy for a given @xmath77 normally occurs along this line .
the saturation point is obtained by minimizing the energy per nucleon @xmath78 , @xmath79 which yields the saturation density @xmath80 ( equivalently , @xmath81 ) and energy @xmath82 .
figure [ fig : nme_m3ya ] illustrates @xmath83 as a function of @xmath77 for the symmetric nuclear matter with the m3y - type as well as with the sly5 and d1s effective interactions .
we set @xmath84 , where @xmath85 ( @xmath86 ) is the measured mass of a proton ( a neutron ) .
the parameters for @xmath31 and @xmath32 of the m3y - type interactions are listed in table [ tab : param_m3y ] .
as mentioned above , the m3y - p0 interaction gives no saturation point .
we do have saturation points in m3y - p1 and m3y - p2 owing to @xmath32 .
differences among the saturating forces , _
i.e. _ sly5 , d1s , m3y - p1 and m3y - p2 , are small at @xmath87 . at relatively high density ( @xmath88 ) , the m3y - p1 and the m3y - p2 interactions have lower @xmath83 than sly5 and higher than d1s .
the values of @xmath81 and @xmath82 are tabulated in table [ tab : nmsat ] .
the m3y - p1 set has been determined so as to give @xmath89 and @xmath90 . in figs .
[ fig : nmv_even ] and [ fig : nmv_odd ] , contribution to @xmath83 from each of the se , te , so and to channels in @xmath35 is shown as a function of @xmath76 .
sum of all these channels and the kinetic energy @xmath91 is equal to @xmath83 in fig .
[ fig : nme_m3ya ] . as is viewed in fig .
[ fig : nmv_even ] , the te channel takes a minimum at @xmath92 except for m3y - p0 and m3y - p1 , primarily responsible for the saturation at @xmath93 . in the d1s interaction ,
the energy out of the se channel monotonically goes down .
this is not compatible with the presence of the strong short - range repulsion in the @xmath0 force , and causes an unphysical property in the neutron matter , as will be shown in section [ sec : nmprop2 ] .
both the so and to channels do not contribute to @xmath83 significantly for @xmath94 ( _ i.e. _ @xmath95 ) . while the so channel becomes attractive and the to channel stays small in the sly5 and the d1s interactions , both channels are repulsive in the m3y - type interactions at @xmath96 , including m3y - p0 .
a certain part of this character of the m3y - type interactions comes from the opep part .
the curvature at the saturation point with respect to @xmath77 is proportional to the incompressibility , @xmath97 the effective mass ( @xmath98-mass ) at the saturation point @xmath99 is defined by @xmath100 the volume asymmetry energy corresponds to the curvature of @xmath83 with respect to @xmath101 , @xmath102 analogously , the following coefficients are defined from the curvatures of @xmath83 with respect to @xmath103 and @xmath104 , @xmath105 these coefficients @xmath106 , @xmath107 and @xmath108 are relevant to the spin and isospin responses in finite nuclei . in table
[ tab : nmsat ] we also compare @xmath109 , @xmath110 , @xmath107 , @xmath106 and @xmath108 among the effective interactions .
the incompressibility @xmath109 is sensitive to @xmath111 in @xmath32 .
the experimental value of @xmath109 has been extracted from the excitation energies of the giant monopole resonances . despite a certain model - dependence ,
most non - relativistic models are consistent with the experiments if @xmath112 . for finite - range interactions , _
i.e. _ the gogny and the m3y - type interactions , @xmath113 seems to give reasonable values of @xmath109 , while in the skyrme interactions @xmath114 looks favorable , because of the momentum - dependent terms in @xmath31 .
the @xmath98-mass is empirically known to be @xmath115 @xcite .
the m3y - type interactions tend to yield slightly smaller @xmath110 than the sly5 and the d1s interactions .
the volume asymmetry energy @xmath107 is important in reproducing global trend of the binding energies for the @xmath116 nuclei . from empirical viewpoints
@xmath117 seems appropriate , as is fulfilled in the m3y - type interactions under consideration .
the @xmath106 and @xmath108 coefficients are relevant to the spin degrees of freedom .
the kinetic energy has a certain contribution to @xmath106 and @xmath108 , as well as to @xmath107 , which amounts to about @xmath118 at @xmath119 equally for @xmath107 , @xmath106 and @xmath108 .
the interaction @xmath35 gives rise to the rest of these coefficients .
both the m3y - type interactions have similar tendency with respect to these coefficients .
it is remarkable that @xmath108 is substantially larger in the m3y - type interactions than @xmath106 .
as is suggested by close @xmath106 and @xmath108 values between m3y - p1 and m3y - p2 , the original m3y interaction already carries this feature .
in particular , the opep part included in the m3y - type interactions plays a significant role , increasing @xmath108 by about @xmath120 . on the other hand , @xmath106 and @xmath108
are comparable in the gogny d1s interaction , and we have even @xmath121 in the skyrme sly5 interaction . in the sly5 case , @xmath108 is close to the value due only to the kinetic energy .
global characters of the spin and isospin responses are customarily discussed in terms of the landau parameters .
formulae on the landau parameters at the zero temperature are given in appendix [ app : landau ] .
we compute the parameters of eq .
( [ eq : landau ] ) .
the results are shown in table [ tab : landau ] .
it is remarked that the m3y - p1 and m3y - p2 interactions give similar results .
the @xmath122 and the @xmath123 parameters are closely related to the @xmath106 and the @xmath108 coefficients , respectively . it has been known that @xmath124 is small , while @xmath125 should be relatively large @xcite . although it is not easy to extract precise values of the landau parameters from experimental data because they could depend on the interaction forms , qualitative trend will not depend on effective interactions .
the m3y - type interactions seem to have reasonable characters on the spin and isospin responses , while sly5 and d1s do not , although the spin and isospin natures of the skyrme interactions seem to be improved if the ls currents are ignored @xcite .
it is likely that the difference in these coefficients may significantly influence predictions of the spin and isospin responses of finite nuclei .
we turn to the asymmetric nuclear matter . in fig . [
fig : nme_m3yb ] , energies per nucleon @xmath83 are depicted as a functions of @xmath77 for the spin - saturated ( _ i.e. _ @xmath126 ) nuclear matter with @xmath127 and @xmath128 .
the results from the m3y - type interactions are compared with those of the skyrme and the gogny interactions .
energies of the spin - saturated neutron matter ( _ i.e. _ @xmath129 ) are presented in fig .
[ fig : nme_m3yc ] . results from a microscopic calculation in ref .
@xcite are also shown as a reference .
although the dependence on the interactions is not strong at low densities even for the neutron matter , it becomes stronger at @xmath130 as @xmath131 increases . in the d1s result for the neutron matter , @xmath83 has a maximum at @xmath132 and goes to @xmath133 as @xmath134 .
this unphysical behavior arises from @xmath135 in the d1s set , which implies no density - dependence in the se channel ( see eq .
( [ eq : ddchannel ] ) ) .
this could also give rise to a problem in practical calculations for finite nuclei . with the sly5 interaction @xmath83
goes up rapidly at any @xmath101 , because of the momentum - dependence of the interaction .
in contrast to them , the m3y - type interactions give moderate @xmath83 for the neutron matter .
the microscopic energy of ref .
@xcite lies between those of m3y - p1 and m3y - p2 . it will be possible , if necessary , to adjust the parameters of the m3y - type interactions to the microscopic results
we next discuss properties of doubly magic nuclei in the hf approximation . in the calculations for finite nuclei
, we use the algorithm presented in ref .
@xcite , where the following s.p .
bases are employed , @xmath136^{(j)}_m\ , ; \nonumber\\ r_{\alpha\ell j}(r ) & = & \mathcal{n}_{\alpha\ell j}\ , r^{\ell+2p_\alpha}\exp[-(r/\nu_\alpha)^2]\ , .
\label{eq : basis } \end{aligned}\ ] ] here @xmath137 expresses the spherical harmonics .
we drop the isospin index without confusion .
the index @xmath111 indicates @xmath138 ( a non - negative integer ) and @xmath139 , simultaneously .
by choosing @xmath138 and @xmath139 appropriately , these bases span the space equivalent to that of the harmonic - oscillator ( ho ) bases , as well as they can form the kamimura - gauss ( kg ) basis - set @xcite . without parameters specific to mass number or nuclide such as @xmath140 , a single set of the kg bases is applicable to wide range of nuclides . in the following calculations we apply the hybrid basis - set @xcite for the nuclei with @xmath141 , in which an ho basis is added to the kg basis - set , while the ho basis - set with @xmath142 and @xmath143 for heavier nuclei . in finite nuclei
the non - central forces are important as well . in the m3y interaction , the ls force @xmath33 and the tensor force @xmath34
are taken by setting @xmath144 and @xmath145 in eq .
( [ eq : effint ] ) .
we here fix the range parameters as in @xmath31 ; @xmath146 , @xmath147 for @xmath33 , and @xmath148 , @xmath149 for @xmath34 .
the coupling constants in the m3y - p2 set are tabulated in table [ tab : param_m3yb ] , together with those in the original m3y - p0 set . in m3y - p2 , the enhancement factor for @xmath33 is taken to be 1.8 and the reduction factor for @xmath34 to be 0.12 .
the binding energies and the rms matter radii obtained from the hf calculations with m3y - p2 are shown in table [ tab : dmprop ] , in comparison with those of the sly5 and the d1s interactions , as well as with the experimental data .
the one - body terms of the center - of - mass ( c.m . ) energy are removed before iteration .
the contribution of the two - body terms is subtracted from the convergent hf wave - functions , in the d1s and the m3y - p2 results .
there are also spurious c.m .
effects in the matter radii , @xmath150\ , .
\label{eq : rcm}\end{aligned}\ ] ] the first term in the right - hand side is expressed by one - body operators with a correction factor @xmath151 .
we need two - body operators for the second term . for the d1s and the m3y - p2 interactions we
fully remove the c.m .
contribution according to eq .
( [ eq : rcm ] ) . for the sly5 interaction
we use only the one - body terms with the correction factor , ignoring the two - body terms in eq .
( [ eq : rcm ] ) , as in calculating the energies .
wave - functions of the doubly magic nuclei are considered to be well approximated in the spherical hf approaches .
it should still be noted that correlations due to the residual interaction could influence their properties .
therefore we do not pursue fine tuning of the parameters . as shown in table [
tab : dmprop ] , the m3y - p2 set is fixed so as to reproduce the measured binding energies of the doubly magic nuclei , including @xmath152zr , within about 5mev accuracy .
the binding energies of these nuclei obtained from the sly5 and the d1s interactions are in agreement with the experimental data within 3mev , slightly better than m3y - p2 .
we do not have to take this difference seriously , before evaluating influence of the residual interactions . as well as the binding energies ,
the rms matter radii of these nuclei are reproduced by the m3y - p2 set similarly well to the other available interactions . in table
[ tab : ls ] we present the neutron s.p .
energies @xmath153 and @xmath154 around @xmath155o . the enhancement factor for @xmath33 in the m3y - p2 set has been adjusted approximately to the experimental value of this s.p .
energy difference .
the reduction factor for @xmath34 has been determined so as to reproduce the s.p .
energy ordering for @xmath156pb . without this reduction factor ,
the orbits with higher @xmath157 have too high energies .
the resultant s.p .
levels in @xmath156pb with m3y - p2 are depicted in fig .
[ fig : pb_spe ] . the levels obtained from d1s and the experimental s.p .
levels are also shown .
the overall level spacings are related to @xmath99 shown in table [ tab : nmsat ] . in the usual hf calculations the level spacings tend to be larger than the observed ones , and it is not ( and should not be ) remedied until the correlations due to the residual interaction ( or the @xmath158-mass ) are taken into account @xcite .
this is also true in the present case .
we find that m3y - p2 yields as plausible s.p .
levels as d1s does .
we thus confirm that the m3y - p2 interaction well describes global nature of stable nuclei .
in the preceding section we have shown that the m3y - p2 interaction reproduces the properties of the doubly magic nuclei to a similar accuracy to the sly5 and the d1s interactions . at a glance , the spin - isospin characters in the nuclear matter , which have been discussed in section [ sec : nmprop ] via @xmath108 and @xmath123 , do not seem to influence the nuclear properties around the ground states .
however , the spin and isospin characters influence s.p .
energies of finite nuclei .
thereby they may affect even the ground state properties . in this section
we illustrate this point by the neutron orbits in the @xmath159 isotones , following the arguments in ref .
@xcite , although precise studies in this line are beyond scope of this paper . as was suggested in ref .
@xcite , the proton - number ( @xmath160 ) dependence of the neutron s.p .
energy @xmath161 relative to @xmath162 can sizably be affected by effective interactions .
figure [ fig : dspe_n16 ] depicts @xmath163 obtained from the spherical hf calculations in the @xmath159 isotones .
though it is not obvious whether the ground states of all of these isotones are well approximated by the spherical hf wave - functions , it is meaningful to see the s.p .
energies , which often give an indication to magic or submagic numbers . for d1s
we reduce the number of bases in eq .
( [ eq : basis ] ) to avoid instability occurring for some @xmath159 nuclei , which probably relates to the unphysical behavior in the neutron matter .
it is found that , if viewed as a function of @xmath160 , @xmath164 strikingly depends on the interactions . with the m3y - p2 interaction , @xmath164 increases as @xmath160 goes from @xmath165 to @xmath166 .
we have confirmed @xcite that even m3y - p1 ( with appropriate @xmath33 and @xmath34 ) shows similar behavior and that a significant part of this feature originates in the opep part in @xmath31 .
it is thus suggested that this behavior of @xmath164 is correlated to the spin - isospin property in the nuclear matter . for comparison
, we also show the s.p .
energies obtained from the reliable shell model interaction for the @xmath167-shell nuclei , the so - called usd interaction @xcite . for this purpose we define effective values of the s.p .
energies for each nucleus @xmath168 from the shell model space and interaction , which correspond to those of the spherical hf calculations , as @xmath169 where the sum with respect to @xmath170 runs over the valence orbits . for @xmath171 , we assume that the nucleons occupy the s.p .
orbits from the bottom , according to @xmath172 . from these s.p .
energies we obtain @xmath173 for individual nucleus .
this definition is equivalent to the effective s.p .
energies in ref .
@xcite for the @xmath174 nuclei .
the @xmath175 values are also shown in fig .
[ fig : dspe_n16 ] .
it is noted that in the shell model approaches the nucleus - dependence of the s.p .
wave - functions is not fully taken into account .
effects of rearrangement in the wave - functions of the deeply bound orbits are renormalized into the interactions among the valence nucleons .
in contrast , in the hf approaches the s.p .
wave - functions are determined self - consistently , from nucleus to nucleus
. therefore the shell model s.p .
energies do not agree with their hf counterparts
. however , there should be qualitative correspondence , which arises from basic characters of the effective interactions .
it is remarked that the m3y - p2 interaction has the same trend of @xmath164 , in terms of the @xmath160-dependence , as the usd interaction .
it has been suggested @xcite that the interaction in the @xmath176 channel , which will be linked to @xmath108 or to @xmath123 , is significant to the magic numbers in highly neutron - rich nuclei , and that the @xmath160-dependence of the s.p .
energies in this region could be relevant to the new magic number @xmath159 @xcite .
the present results are fully consistent with the arguments in ref .
@xcite , although we can not draw conclusions on the magic number problem without assessing influence of the residual interactions .
we have developed effective interactions to describe low energy phenomena of nuclei .
starting from the m3y interaction , we introduce a density - dependent contact term and modify several parameters in a phenomenological manner , whereas maintaining the opep part in the central force . in order to view basic characters of the interactions ,
the hartree - fock calculations are implemented for the infinite nuclear matter ( for which useful formulae are newly derived ) and for several doubly magic nuclei .
we have shown that a parameter - set called m3y - p2 describes their properties plausibly .
the properties which are well treated by the skyrme sly5 and/or the gogny d1s interactions are also reproduced by the m3y - p2 interaction . however , a remarkable difference is found in the properties relevant to the spin degrees of freedom in the nuclear matter .
the m3y - type interactions seem to give reasonable spin and isospin properties , in which the opep part contained in @xmath31 plays a significant role . we have also shown that the difference in the spin - isospin property affects the s.p .
energies in finite nuclei to a considerable extent
. it will be interesting to apply extensively the m3y - type interactions , particularly to the magic number problems far from the @xmath1-stability .
although the m3y - p2 interaction seems to have various desired characters , there still remains a certain room for further tuning of the parameters .
it should be noted that this parameter - set will not be a unique choice to reproduce the properties of the nuclear matter and the doubly magic nuclei .
effective interaction might not be constrained sufficiently only from the hf calculations .
the pairing effects in nuclei give valuable information on the effective interaction , primarily on the se channel .
comparison of the matrix elements with reliable shell model interactions will also be helpful , if the core polarization effects are treated appropriately .
these points will be discussed in future publications .
.parameters of central forces ( including @xmath32 ) in the original and modified m3y interactions .
see text for the @xmath177 parameters .
[ tab : param_m3y ] [ cols="^,^,>,<,>,<,>,<",options="header " , ] in the symmetric nuclear matter for several effective interactions
. the thick dotted , dot - dashed and solid lines represent the results with the m3y - p0 , m3y - p1 and m3y - p2 interactions , respectively , while the thin dashed and solid lines those with the sly5 and d1s interactions .
[ fig : nme_m3ya],height=377 ] .
see fig .
[ fig : nme_m3ya ] for conventions .
[ fig : nmv_even],height=604 ] . in both channels ,
the results of m3y - p0 is equal to those of m3y - p1 , which are presented by the dot - dashed line .
see fig .
[ fig : nme_m3ya ] for the other conventions .
[ fig : nmv_odd],height=604 ] in the asymmetric nuclear matters with @xmath127 and @xmath128 for several effective interactions .
see fig .
[ fig : nme_m3ya ] for conventions .
[ fig : nme_m3yb],height=604 ] in the neutron matter for several effective interactions .
the circles are the results of ref .
see fig .
[ fig : nme_m3ya ] for the other conventions .
[ fig : nme_m3yc],height=340 ] pb .
experimental values are extracted from refs .
[ fig : pb_spe],height=453 ] for the @xmath159 isotones .
the thick solid , dotted , thin solid and dashed lines correspond to the results with the m3y - p2 , usd , d1s and sly5 interactions , respectively . [
fig : dspe_n16],height=377 ]
in this appendix we derive formulae concerning the interaction part of eq .
( [ eq : nme ] ) .
the form of eq .
( [ eq : effint ] ) is assumed for @xmath178 .
each term of @xmath31 is expressed as @xmath179 .
its non - antisymmetrized matrix element in the plane wave states of eq .
( [ eq : nm - spwf ] ) is evaluated as @xmath180 where @xmath181 , @xmath182 , @xmath183 , @xmath184 , @xmath185 , @xmath186 and @xmath187 is the fourier transform of @xmath188 , @xmath189 the density - dependent interaction @xmath32 is also handled in a similar manner , since the density behaves like a constant in the nuclear matter . for the hartree term
we have @xmath190 and @xmath191 , while @xmath192 and @xmath193 for the fock term .
therefore both terms satisfy @xmath194 . for the relative momentum the hartree term ( the fock term ) yields @xmath195 ( @xmath196 ) .
contribution of the two - body interaction to the nuclear matter energy is obtained by integrating @xmath197 in eq .
( [ eq : nme - tbme ] ) up to the fermi momenta .
we here consider general cases where the fermi momentum may depend on spin and isospin . in order to take into account the spin - isospin dependence
, we integrate @xmath197 in the range @xmath198 and @xmath199 .
the integration is immediately carried out for the hartree term , as far as @xmath200 is momentum - independent , since the integrand depends neither on @xmath201 nor on @xmath202 , @xmath203 for the fock term contribution , the integral with respect to @xmath201 and @xmath202 is converted to the one with respect to @xmath204 and @xmath205 .
we here assume @xmath206 without loss of generality , owing to the symmetry @xmath207 . handling the range of integral carefully
, we obtain the following expression , @xmath208\end{aligned}\ ] ] these formulae are general to multi - component uniform fermi liquids with equal masses .
in handling the spin - isospin degrees of freedom , we rewrite the central force in eq .
( [ eq : effint ] ) as @xmath209 the relations between the coupling constants are @xmath210 after summing over the spin - isospin degrees of freedom , the interaction energy is given by @xmath211\ , .
\label{eq : nme - tot}\end{aligned}\ ] ] in eq .
( [ eq : nme - tot ] ) we regard the sum over @xmath19 to include @xmath32 . it is noted that @xmath212 , which is used to obtain the energy per nucleon @xmath83 .
we next calculate the @xmath213 functions for typical interaction forms . 1 . _
@xmath214 interaction _
+ if @xmath215 , @xmath216 and therefore we have @xmath217 2 . _ @xmath77-dependent @xmath214 interaction _ + since the density is a constant in the uniform nuclear matter , the @xmath213 functions for @xmath218 are similar to the above case , @xmath219 note that @xmath77 is a function of the fermi momenta , when we take derivatives of the @xmath213 functions
_ gauss interaction _
+ for @xmath220 , we have @xmath221 , deriving @xmath222 and @xmath223\ , , \nonumber\\ \label{eq : w - g}\end{aligned}\ ] ] where @xmath224 in eq .
( [ eq : w - g ] ) we have postulated @xmath225 again .
yukawa interaction _
+ for the yukawa interaction we set @xmath226 , leading to @xmath227 .
this yields @xmath228 and @xmath229\,.\end{aligned}\ ] ] 5 .
_ momentum - dependent @xmath214 interaction _ + in the skyrme interaction we have momentum - dependent terms with the form @xmath230 and @xmath231 .
the former operates only on the even channels and yields @xmath232 the latter acts on the odd channels , giving @xmath233 the incompressibility @xmath109 and the spin - isospin curvatures @xmath107 , @xmath106 , @xmath108 are expressed by the derivatives of the @xmath213 functions . the single - particle energy @xmath234 defined in eq .
( [ eq : nm - spe ] ) is also expressed by the derivative of the @xmath213 functions .
we first rewrite the integral in eq .
( [ eq : nme ] ) as @xmath235 this immediately gives @xmath236 therefore , @xmath237\ , , \nonumber\\\end{aligned}\ ] ] where we use the short - hand notation @xmath238 it is now obvious that the effective mass of eq .
( [ eq : m * ] ) is expressed by using the second derivative of the @xmath213 functions .
let us denote the occupation probability of the s.p .
states of eq .
( [ eq : nm - spwf ] ) by @xmath239 . the nuclear matter energy of eq .
( [ eq : nme - tot ] ) can be rewritten as @xmath240 the landau coefficient is defined by @xmath241 for the interaction independent of momentum and of density , it is straightforward to write down the coefficients of eq .
( [ eq : def - lan ] ) in terms of @xmath197 , within the hf theory at the zero temperature .
noticing that @xmath77 also depends on @xmath239 , we evaluate contribution of the density - dependent @xmath214 interaction @xmath242 to @xmath243 as @xmath244\,,\end{aligned}\ ] ] where @xmath245 and @xmath246 .
apart from the spin and isospin degrees of freedom , the momentum - dependent @xmath214 interactions @xmath230 and @xmath231 contribute to @xmath243 by @xmath247 in characterizing effective interactions , we view the landau coefficients for the symmetric nuclear matter , where @xmath248 for any @xmath68 and @xmath67 . while formulae for the landau parameters were derived for the skyrme interaction in ref .
@xcite and for the gogny interaction in ref .
@xcite , we here derive expressions for interactions with the form of eq .
( [ eq : effint ] ) in more general manner .
it is customary to transform the @xmath249 variables into the following ones , @xmath250 since @xmath251 , all the off - diagonal coefficients with respect to @xmath252 vanish .
the diagonal coefficients are redefined as @xmath253 the hartree terms of the momentum- and density - independent interactions yield @xmath254 while the fock terms @xmath255 where @xmath256 contribution of the density - dependent interaction @xmath257 is given by @xmath258 for momentum - independent interactions such as the gogny interaction and the m3y - type interactions , the landau coefficients are obtained by @xmath259 , and so forth .
the momentum - dependent @xmath214 interactions yield @xmath260 where the upper row corresponds to the even channel interaction @xmath261 , while the lower to the odd channel interaction @xmath262 , respectively .
equation ( [ eq : lan - md ] ) is available for the skyrme interactions in which the ls currents are not ignored . 1 . _
@xmath214 interaction _ + substituting @xmath264 by @xmath265 , we obtain @xmath266 2 .
_ gauss interaction _ + because @xmath221 , eq .
( [ eq : g - fac ] ) leads to @xmath267 for @xmath268 and @xmath265 , we have @xmath269 3 . _
yukawa interaction _
+ for the yukawa interaction we use @xmath227 . inserting it into eq .
( [ eq : g - fac ] ) , we obtain for even @xmath157 @xmath270 \ , . \nonumber\\\end{aligned}\ ] ] and for odd @xmath157 @xmath271 \ , . \nonumber\\\end{aligned}\ ] ] for @xmath268 and @xmath265 , we have @xmath272 \,.\end{aligned}\ ] ] this work is financially supported as grant - in - aid for scientific research ( c ) , no . 13640263 , by the ministry of education , culture , sports , science and technology , japan .
numerical calculations are performed on hitac sr8000 at institute of media and information technology , chiba university , at information technology center , university of tokyo , and at computing center , hokkaido university .
99 s. c. pieper , k. varga and r. b. wiringa , phys .
c * 66 * , 044310 ( 2002 ) .
i. tanihata , nucl phys . *
a654 * , 235c ( 1999 ) . t.
lett . * 87 * , 082502 ( 2001 ) .
d. vautherin and d. m. brink , phys .
c * 5 * , 626 ( 1972 ) .
j. decharg and d. gogny , phys .
c * 21 * , 1568 ( 1980 ) .
l. coraggio _
et al . _ ,
c * 58 * , 3346 ( 1998 ) ; _ ibid . _ * 60 * , 064306 ( 1999 ) .
j. p. schiffer and w. w. true , rev .
phys . * 48 * , 191 ( 1976 ) .
g. bertsch , j. borysowicz , h. mcmanus and w.g .
love , nucl .
a284 * , 399 ( 1977 ) .
b. a. brown , w. a. richter , r. e. julies and b. h. wildenthal , ann .
( n.y . ) * 182 * , 191 ( 1988 ) .
d. t. khoa , w. von oertzen and a. a. ogloblin , nucl . phys .
* a602 * , 98 ( 1996 ) ; d. t. khoa , g. r. satchler and w. von oertzen , phys .
c * 56 * , 954 ( 1997 ) .
h. nakada and m. sato , nucl . phys .
* a699 * , 511 ( 2002 ) ; _ ibid .
_ * a714 * , 696 ( 2003 ) .
j. w. negele , phys .
c * 1 * , 1260 ( 1970 ) ; x. campi and d. w. sprung , nucl .
a194 * , 401 ( 1972 ) .
f. hofmann and h. lenske , phys .
c * 57 * , 2281 ( 1998 ) .
h. a. bethe , ann .
* 21 * , 93 ( 1971 ) .
n. anantaraman , h. toki and g. f. bertsch , nucl
a398 * , 269 ( 1983 ) .
k. and and h. band , prog .
. phys . * 66 * , 227 ( 1981 ) ; k. suzuki , r. okamoto and h. kumagai , phys .
c * 36 * , 804 ( 1987 ) ; s. c. pieper and v. r. pandharipande , phys .
rev . lett . * 70 * , 2541 ( 1993 ) .
j. f. berger , m. girod and d. gogny , comp .
. comm . * 63 * , 365 ( 1991 ) .
e. chabanat , p. bonche , p. haensel , j. meyer and r. schaeffer , nucl .
a635 * , 231 ( 1998 ) . c. mahaux , p. f. bortignon , r. a. broglia and c. h. dasso , phys
. rep . * 120 * , 1 ( 1985 ) . c. gaarde
_ et al . _ ,
* a369 * , 258 ( 1981 ) ; t. suzuki , nucl .
a379 * , 110 ( 1982 ) ; g. bertsch , d. cha and h. toki , phys .
c * 24 * , 533 ( 1981 ) .
m. bender , j. dobaczewski , j. engel and w. nazarewicz , phys . rev .
c * 65 * , 054322 ( 2002 ) .
b. friedman and v. r. pandharipande , nucl .
a * 361 * , 502 ( 1981 ) .
h. kameyama , m. kamimura and y. fukushima , phys .
c * 40 * , 974 ( 1989 ) .
g. audi and a. h. wapstra , nucl . phys . *
a595 * , 409 ( 1995 ) .
d. t. khoa , h. s. than and m. grasso , nucl .
a722 * , to be published .
g. d. alkhazov , s. l. belostotsky and a. a. vorobyov , phys .
42 * , 89 ( 1978 ) .
r. b. firestone _
_ , _ table of isotopes _ , 8th edition ( john wiley & sons , new york , 1996 ) .
h. nakada , nucl . phys . *
a722 * , 117c ( 2003 ) .
a. ozawa , t. kobayashi , t. suzuki , k. yoshida and i. tanihata , phys .
* 84 * ( 2000 ) 5493 .
j. ventura , a. polls , x. vias and e. s. hernandez , nucl . phys .
* a578 * , 147 ( 1994 ) . | by introducing a density - dependent contact term , m3y - type interactions applicable to the hartree - fock calculations are developed . in order to view basic characters of the interactions ,
we carry out calculations on the uniform nuclear matter as well as on several doubly magic nuclei .
it is shown that a parameter - set called m3y - p2 describes various properties similarly well to the skyrme sly5 and/or the gogny d1s interactions . a remarkable difference from the sly5 and the d1s interactions
is found in the spin - isospin properties in the nuclear matter , to which the one - pion - exchange potential gives a significant contribution . affecting the single - particle energies
, this difference may play a certain role in the new magic numbers in unstable nuclei . |
entropy is a measure of the _ size _ of a data distribution contained within a bounded region ( distribution support ) of some space . in a thermodynamic context
this distribution size is interpreted as the number of quantum states accessible to a dynamical system given macroscopic constraints . more generally ,
if a measure space is partitioned , the measure distribution size is estimated by the effective number of partition elements given the distribution weighting .
definition of the space partition is a central element of entropy calculation .
the partition is sometimes defined as a small - scale limiting partition of the space ( _ e.g. , _ thermodynamic limit , limit procedures in classical analysis @xcite ) , sometimes based on properties of the data distribution itself and/or on the analysis goals ( as in wavelet analysis @xcite ) .
we define entropy as an explicit function of the partition definition , a scaled binning of the measure space .
we calculate the entropy and related quantities as functions of the partition scale(s ) , similar to the multi - scale partition approaches used in fractal dimension calculations @xcite and image deconvolution @xcite . unlike some analysis methods we invoke no scale limits .
quantities are defined on bounded scale intervals explictly excluding asymptotic limits ; the analysis system is in this sense _ scale local_. the end result of this approach is an entropy which represents _ arbitrary _ data correlations as a distribution on scale , particularly useful for problems where the detailed scaling behavior of correlations over substantial scale intervals is of interest ( _ e.g. , _ condensation , coalescence , critical phenomena , strange attractors ) , where instrumental effects may distort scale distributions , and where the correlation structure is not simply expressible as a power law or other elementary function .
@xcite in this paper we describe precision binning methods , define the basic scale - local entropy measure and generalize other aspects of information theory to define the scale - dependent entropy difference or information between an object distribution and a model reference as a differential correlation measure . based on scale derivatives of entropy and information
we define scale - local dimension and dimension transport as generalizations of conventional counterparts based on limit concepts .
we apply these correlation measures to several simulations and real data analysis problems .
the entropy definition employed here is based on @xmath1 , the rank-@xmath2 correlation integral at scale @xmath3 .
given a data distribution in a @xmath0-dimensional primary measure space spanned by variables @xmath4 , we consider a set of corresponding correlation spaces containing all possible @xmath2-point clusters of data points ( @xmath2-tuples ) .
there is one such @xmath2-point distribution ( in a @xmath2-fold cartesian product space ) for each unique @xmath2 value .
the @xmath2-point correlation integral is the projection of the @xmath2-point distribution onto its difference subspace spanned by @xmath5 , integrating over the sum variable(s ) .
@xcite the integration limit of the correlation integral on the difference variables is in the simplest case ( isotropic binning ) the single scale of the analysis .
the reciprocal of the correlation integral estimates the _ effective _ bin number in the @xmath6-dimensional difference subspace .
it s counterpart in the primary measure space is the @xmath7 root , the effective bin number in the primary space .
defining entropy as the logarithm of the effective primary - space bin number is consistent with entropy as a logarithmic size measure and is a generalization of the thermodynamic definition , the logarithm of the number of accessible states .
the correlation integral can be approximated by binning the primary measure space , and expressed in terms of normalized bin contents @xmath8 , in which case @xmath9 .
this results in the rank-@xmath2 rnyi entropy @xcite @xmath10 \simeq \frac{1}{1-q}\log \left[\sum_{i=1}^{m(e)}p_{i}(e)^q\right].\ ] ] the entropy @xmath11 , the number of occupied bins @xmath12 , and the bin probability @xmath13 are explicit functions of the binning scale @xmath3 .
more generally , a non - isotropic binning ( one utilizing bins without unit aspect ratio ) would imply a multidimensional scale space .
given scale - local entropy we define scale - local information as a basis for differential comparisons between object and reference distributions , data and model .
there are a number of possible information definitions from information theory and topology , with significant differences in performance .
we define information here as the difference between entropies for object and reference distributions .
this implies that the effective bin number for an object distribution is compared _ in ratio _ to that of a reference .
nonzero information implies _ multiplicative _ reduction of effective bin number by increased correlation structure in the object distribution relative to the reference ( cumulant analysis is an alternative differential approach emphasizing _ linear or additive _ reduction of distribution size ) .
scale - local information is then defined as @xmath14 information so defined provides a differential comparison between a data or _
object _ distribution and a model or _
reference_. for example , the reference distribution for an arbitrary point set would be a distribution with the same number of points which maximally ` fills ' the bounded support a uniform random distribution . since the uniform reference is useful in many applications we derive explicit analytic forms for its entropy and dimension .
although these scale - local analysis methods are completely general as to the nature of the measure distribution , for the purpose of illustration we emphasize point sets in this paper .
the measure distribution is then a set of points in a @xmath0-dimensional embedding space ( _ e.g. , _ the distribution of particles from a heavy - ion collision in momentum space ) .
the analysis begins by applying a partition to the embedding space . for algebraic simplicity
we consider a grid of @xmath0-dimensional cubes , an isotropic binning . at each scale
there is a continuum of possibilities for the relative position of the binning system on the embedding space .
differing partition placement effects the analysis in general , and there is no _ a priori _ reason to prefer any single placement .
thus , we average ( dither ) over all partition placements to calculate the entropy at each scale .
we define a dithering phase @xmath15 for each of the @xmath0 embedding - space dimensions .
the relationship between the partition system and the measure distribution is controlled by varying @xmath15 . to implement dithering
we calculate the correlation integral @xmath16 of an event @xmath17 times at each scale , incrementing @xmath15 each time .
finally we average over these results to obtain the entropy .
the dithered entropy @xmath18 is thus @xmath19 .
\end{aligned}\ ] ] where @xmath20 is the bin probability of the @xmath21th bin for binning phase @xmath15 and scale @xmath3 .
@xmath13 is normalized so that @xmath22 at each scale @xmath3 and dithering phase @xmath15 , the sum taken over all occupied bins in each partition .
a simple application of scale - local entropy and dithered binning illustrates the analysis process .
we obtain the scaled entropy of a 2d uniform distribution of @xmath23 randomly generated points on a unit - square support for several values of index @xmath2 .
a monte carlo simulation with 50k points is shown in figure [ 1drgud ] along with analysis results for @xmath24 . .
a box plot of the distribution itself is shown in the left panel , the right panel shows the measured entropy for @xmath25 ( dashed line ) , @xmath26 ( dotted line ) , and @xmath27 ( dot - dashed line).,width=384 ] to interpret these results we consider small- , intermediate- , and large - scale regimes . in the small - scale limit , well below the point - separation scale ( @xmath28 ) , each distribution point occupies a single bin ; there are n occupied bins ( @xmath29 ) , each with bin probability @xmath30 .
thus , at small scale the entropy approaches @xmath31=\log n.\ ] ] at intermediate scales we idealize the uniform point set to a continuum ( @xmath32 ) .
the number of occupied bins is then simply @xmath33 and the bin probability is @xmath34 giving @xmath35 . at scales substantially greater than the boundary scale ( @xmath36 ) the entire distribution is contained in a single bin : @xmath37 and @xmath38 .
the rank-@xmath2 entropy thus vanishes at scales much larger than the distribution boundary size .
the detailed @xmath2-dependence near the particle - separation and boundary scales is amenable to an analytic treatment .
information is a relative quantity , a matter of definition .
it is impossible to make an absolute determination of the information content of an arbitrary distribution . using a maximum - entropy reference we can measure information relative to a distribution which is minimally correlated ( given certain constraints ) .
this motivates us to derive an algebraic form for the scale - local entropy of a _ bounded uniform distribution _ ( bud ) .
this distribution represents a maximum - entropy hypothesis within a boundary , a maximum filling of the distribution support .
this is a correlation reference from which any object distribution may deviate with reduced entropy .
the derivation is presented in two parts : scales below and above the boundary scale . for partitions below the boundary scale
there is at least one bin in the interior of the embedding space . to derive an analytic form for the entropy of a bud we consider a two - dimensional distribution ( and
later generalize to @xmath0 dimensions ) .
the bud is defined on a square support with side length @xmath40 . because the distribution is uniform the probability of finding a point in any given bin
is simply determined by the bin area .
figure [ dithervariables2d ] shows a bud binned with a general rectangular binning ( thin dark lines ) .
we calculate the bin contents for each bin ( as a function of scale ) and integrate over all possible dithering configurations to determine the analytic form of the entropy . for bins that are entirely contained within the embedding space ( interior bins )
the bin probability is trivial : @xmath41 is the area of the bin divided by the total area of the support , independent of bin dithering . for edge and corner bins
the problem is more complicated .
we consider each bin type interior ( white ) , edge ( light grey ) and corner ( dark grey ) separately .
corner bins contain both an x- and y - axis support edge .
for all dithering configurations there are four corner bins .
the effective area of these corner bins is @xmath42 scaled down by the fraction of the bin along the @xmath21th axis ( @xmath43 $ ] ) that overlaps the data . by definition ,
the phase along an axis is @xmath44 , where @xmath45 is the bin overlap fraction for the first bin on axis @xmath21 ( @xmath46 $ ] ) .
we define @xmath47 and express the amount of overlap between the last bin on axis @xmath21 and the edge of the support along that axis as @xmath48 . with these definitions , calculating the contribution to the correlation integral of the corner bins is a matter of integrating over all @xmath15 values using the relevant bin probabilities .
labeling the corner bins from right to left starting with the upper left bin we write down the @xmath15-dependent corner bin probabilities as @xmath49^q\left ( \frac{e_{\text{x}}e_{\text{y}}}{l^2}\right)^q \\ \nonumber p_{\text{c}}^q&=&[\phi_{\text{x}}+\delta_{\text{x}}-\text{int } ( \phi_{\text{x}}+\delta_{\text{x}})]^q(1-\phi_{\text{y}})^q\left ( \frac{e_{\text{x}}e_{\text{y}}}{l^2}\right)^q \\ \nonumber p_{\text{d}}^q&=&[\phi_{\text{x}}+\delta_{\text{x}}-\text{int } ( \phi_{\text{x}}+\delta_{\text{x}})]^q[\phi_{\text{y}}+\delta_{\text{y}}-\text{int } ( \phi_{\text{y}}+\delta_{\text{y}})]^q\left ( \frac{e_{\text{x}}e_{\text{y}}}{l^2}\right)^q.\end{aligned}\ ] ] we calculate the dither - averaged correlation integral @xmath50 by integrating over the different dithering configurations for each bin and summing @xmath51 . following this approach
we integrate the @xmath52 expressions over the two dithering variables and sum results to calculate the correlation integral .
the @xmath15 integral over the first corner bin yields @xmath53 the second term is @xmath54^q\ , \rmd\phi_{\text{x}}\ , \rmd\phi_{\text{y } } \\ \nonumber & = & \left(\frac{e_{\text{x}}e_{\text{y}}}{l^2}\right)^q\left(\frac{1}{1+q}\right)\left[\int_{\delta_{\text{y}}}^1 v^q\ , \rmd v+\int_0^{\delta_{\text{y } } } v^q\ , \rmd v\right ] \\
\nonumber & = & \left(\frac{e_{\text{x}}e_{\text{y}}}{l^2}\right)^q\left(\frac{1}{1+q}\right)^2.\end{aligned}\ ] ] the third and fourth terms are similar to the first and second ; each of the four corner bins contributes an @xmath55 term to the total dither - averaged correlation integral .
the contributions from edge bins are simpler to calculate .
the overlap fraction is unity in the direction parallel to the support edge ; along that axis we merely count the number of edge bins .
the integral along the second axis is similar to the corner bins .
again there are four terms , but symmetry simplifies the problem @xmath56 ( 1-\phi_{\text{y}})^q\ , \rmd\phi_{\text{x}}\ , \rmd\phi_{\text{y } } \\ \nonumber & = & \left(\frac{e_{\text{x}}e_{\text{y}}}{l^2}\right)^q \int_0 ^ 1\int_0 ^ 1 \left[\text{int}\left(\frac{l}{e_{\text{x}}}+1+\phi_{\text{x}}\right ) -2\right ] [ \phi_{\text{y}}+\delta_{\text{y}}-\text{int}(\phi_{\text{y}}+\delta_{\text{y}})]^q \ , \rmd\phi_{\text{x}}\ , \rmd\phi_{\text{y } } \\ \nonumber & = & \left(\frac{e_{\text{x}}e_{\text{y}}}{l^2}\right)^q\left(\frac{l}{e_{\text{x}}}-1\right)\left(\frac{1}{1+q}\right).\end{aligned}\ ] ] which is the expression for the x - axis border bins .
contributions from the y - axis bins are obtained by switching indices .
the contribution from edge bins is thus @xmath57.\ ] ] there remains the integral over interior bins , a simple matter of bin counting @xmath58 the full correlation integral can be assembled from the corner , edge , and interior bin integrals @xmath59 \\
\nonumber \fl \;\;\;\;\;\;\;\;\;\ , = \left(\frac{e_{\text{x}}e_{\text{y}}}{l^2}\right)^{q-1 } \left[1+\left(\frac{1-q}{1+q}\right)\left(\frac{e_{\text{x}}}{l}\right)\right]\left[1+\left(\frac{1-q}{1+q}\right)\left(\frac{e_{\text{y}}}{l}\right)\right].\end{aligned}\ ] ] inserting this result into the definition of the rank-@xmath2 entropy we find that @xmath60 \\
\nonumber \fl \;\;\;\;\;\;\;\;\;\
, = \frac{q-1}{1-q } \log \left(\frac{e_{\text{x}}e_{\text{y}}}{l^2}\right)+\frac{1}{1-q } \log \left[1+\left(\frac{1-q}{1+q}\right)\left(\frac{e_{\text{x}}}{l}\right)\right]+\frac{1}{1-q } \log \left[1+\left(\frac{1-q}{1+q}\right)\left(\frac{e_{\text{y}}}{l}\right)\right ] \\
\nonumber \fl \;\;\;\;\;\;\;\;\;\ , = \left\{\log\left(\frac{l}{e_{\text{x}}}\right)+\frac{1}{1-q}\log\left[1+\left(\frac{1-q}{1+q}\right)\frac{e_{\text{x}}}{l}\right]\right\}+\left\{\log\left(\frac{l}{e_{\text{y}}}\right)+\frac{1}{1-q}\log\left[1+\left(\frac{1-q}{1+q}\right)\frac{e_{\text{y}}}{l}\right]\right\}.\end{aligned}\ ] ] because of the additivity of entropy under product this entropy expression contains two terms , one for each axis , which suggests the @xmath0-dimensional generalization : adding an equivalent term for each additional dimension .
for the derivation of eq .
( [ genent ] ) we considered contributions from three types of bins : corner , border , and interior . in the 2d derivation
this approach is only valid when both @xmath62 and @xmath63 .
if either of the scales is larger than the support size there are no _ interior _ bins and eq .
( [ genent ] ) is not valid .
thus , to obtain the entropy expression for a bud at large scale consider a single axis ( we now exploit the additivity of entropy with respect to dimension for uncorrelated systems ) with @xmath61 .
when there is a bin edge in the support there are exactly two corner bins ; we express the bin probability of the second bin in terms of the first @xmath64^q \\ \nonumber p_2^q&=&[1-p_1]^q.\end{aligned}\ ] ] when the support fits completely inside a single bin @xmath65 ; the distance from the support edge to the nearest bin edge is larger than the size of the support ( @xmath66 ) .
we now evaluate the relevant integrals @xmath67^q\ , \rmd\alpha=1-\frac{l}{e},\ ] ] @xmath68 since the definition of corner bins is arbitrary , we could relabel and get the same result ; symmetry requires that @xmath69 .
these results can now be assembled to calculate the 1d scaled entropy for scales larger than the support size @xmath70 \\ \nonumber & = & \frac{1}{1-q } \log \left[1-\frac{l}{e}+\frac{l}{e}\left(\frac{2}{1+q}\right)\right ] \\
\nonumber & = & \frac{1}{1-q } \log \left[1+\left(\frac{1-q}{1+q}\right)\frac{l}{e}\right].\end{aligned}\ ] ] combining results , we obtain the rank-@xmath2 scaled entropy for each degree of freedom of a bud over all scales @xmath71,&for $ e \le l$\cr \frac{1}{1-q } \log\left[1+\left(\frac{1-q}{1+q}\right)\frac{l}{e}\right],&for $ e \ge l$.\cr}\ ] ] we have generated the exact expression for the scale - local entropy in the general case of a @xmath0-dimensional bounded uniform distribution .
dashed , @xmath26 dotted , @xmath27 dot - dashed lines ) with analytic brud entropy for @xmath25 ( solid line ) as a reference .
data and reference entropy distributions differ near the mean interparticle spacing scale ( @xmath78 - 2.35 ) as discussed in the text.,width=384 ] this entropy is precise for a discrete , random point set and @xmath25 . for @xmath79 the reference entropy
differs from the entropy for real point sets in a small scale interval near @xmath80 , the typical two - point separation .
this information ( entropy difference ) derives from the fact that the form of rnyi entropy employed here is based on ordinary moments ( averaged powers of @xmath81 ) , whereas the point - set data are poisson distributed .
a correlation integral based on factorial moments would give zero information for uniform ( uncorrelated ) point sets @xcite .
we could conclude that a factorial - moment approximation to the correlation integral should be used for point sets at least , if not other applications .
however , extending our previous observation that boundedness itself is a form of correlation in a self - consistent system we can also view the point set as a result of increasing correlation ( coalescence ) of a continuous distribution at a characteristic scale .
the point set does have correlations additional to the boundedness of the continuous bud .
the apparent discrepancies in the entropies of brud and real points sets ( the nonzero @xmath82 in fig .
[ 50krud ] ) reveal a genuine correlation feature in a discrete point set compared to a continuum .
the rnyi entropies based on ordinary moments are preferred as a _
more general _ formulation applicable to arbitrary measure distributions .
the information corresponding to the difference between factorial moments and ordinary moments for the uniform random point set is meaningful and can be expressed analytically ( the subject of a future paper ) .
having obtained entropy and information as scale - local distributions we can similarly express the dimension of a distribution as a function of scale .
we start with a conventional definition of dimension based on asymptotic limits @xcite @xmath83 = \lim_{e\rightarrow 0 } \biggl [ -\frac{s_q(e)}{\log ( e ) } \biggr].\ ] ] in this approach dimension is a single number defined in the limit of a zero - scale partition .
the slope of the @xmath18 distribution is evaluated in the limit of zero scale .
this definition favors certain specific types of correlation ( power laws ) , and assumes that the limit exists , which may not be true even in principle . for more general cases
the results can be misleading .
we relax the asymptotic limit restriction as we did with scale - local entropy to obtain a more general _ scale - local _ dimension @xmath84
= -\frac{\partial [ s_q(e)]}{\partial [ \log ( e)]}.\end{aligned}\ ] ] applying this definition to the generalized entropy of a brud in eq .
( [ point - ent ] ) yields @xmath85 dashed , @xmath26 dotted , @xmath27 dot - dashed lines ) .
the solid lines shows the analytical @xmath25 results derived for a generalized brud.,width=384 ] dimension expressed as a scale - local distribution is a novel aspect of this entropy treatment . to understand how scale can affect
the inferred dimension of an object consider the apparent dimension of a planet in the solar system from different viewpoints .
for an observer on one planet other planets appear to the eye as isolated points ( with zero dimension ) . with a powerful telescope
the resolution size ( scale ) of the observation decreases substantially .
planets appear as 2d disks with 1d border .
a radar probe orbiting a planet can determine that the planet at smaller scale has a rich 3d surface and internal structure that could not be supported by a 1d point or 2d plane .
continuing to the atomic scale planets are made of atoms and molecules that appear point - like . at this scale a planet
s dimensionality returns to zero .
this general principle is illustrated in figure [ dq ] .
as an exercise in precision correlation analysis using scale - local entropy and dimension we model cluster formation via condensation by generating a hierarchical point distribution with correlation features distributed over a range of scales .
this model is relevant to phase transitions and complex systems analysis . to create a two - dimensional , two - tier cluster hierarchy on a square region of side @xmath40
we generate a uniform random distribution of @xmath86 cluster sites , providing correlations at the characteristic length scale @xmath87 the mean site separation . at each cluster site
we throw a randomly generated uniform distribution of @xmath88 points with width @xmath89 , giving the distribution a second characteristic length scale @xmath90 . , @xmath91 & @xmath92 .
the scaled dimension of the data ( dashed line ) is compared to the analytic brud reference with the corresponding number of points ( @xmath93 ) at scale @xmath94 ( black solid line ) as well as the analytic reference for @xmath95 at scales @xmath96 and @xmath97 ( gray solid lines).,width=384 ] if the two tiers of the hierarchy are sufficiently separated on scale , as in fig .
[ 2d2steph ] , the sub - structure of the clusters is not evident to the analysis at large scale ( @xmath98 ) .
the distribution appears to be a brud of @xmath86 random points . at smaller scale ( @xmath99 )
the apparent structure is dominated instead by the internal cluster structure , a brud of @xmath88 points .
this two - tiered hierarchical distribution is a first approximation to a self - similar distribution : it appears as a brud at scales @xmath40 and @xmath89 simultaneously ( see fig .
[ 2d2steph ] ) . extending
the hierarchy by recursive self - similar cluster generation would converge to the limiting case of a fractal point distribution over an arbitrarily large scale interval ( _ e.g. , _ cantor set ) .
the ` fractal ' dimension would depend on the relations among the hierarchy scale separation , the cluster size and the point count .
the lower right panel of figure [ 2d2steph ] shows the _ dimension transport _ for the two - tier hierarchy .
dimension transport is defined as the scale derivative of information , @xmath100/\partial [ \log ( e)]$ ] , and is a measure of the scale - dependent dimension difference between reference and object distributions .
dimension transport measures increasing correlation as the transport of dimensionality from larger to smaller scale . in the example of the two - tier hierarchy correlation ( relative to the brud of the same multiplicity )
anticorrelation of points at larger scale is achieved when points condense toward the cluster sites .
the anti - correlation at larger scale results in a reduction of larger - scale dimensionality ( the system appears more point - like ) and an increase in the local point density and dimensionality at smaller scale . ) in a system with a fixed multiplicity ( @xmath101 ) .
data are compared to a reference brud with equal multiplicity.,width=384 ] the extended two - tier condensation example in figure [ 6panelh ] shows how small - scale correlations increase by condensing points of a brud onto cluster sites . at the onset of cluster formation ( @xmath1023000 cluster sites , @xmath1023 points per cluster )
the transport of dimension to smaller scale is barely visible ( but still non - statistical ) .
when the size of the clusters becomes significant ( @xmath1021000 cluster sites , @xmath10210 points per cluster ) the analysis indicates what the eye perceives directly , that the distribution of points is in some way correlated .
when the cluster size is 10% of the number of clusters ( @xmath102300 cluster sites , @xmath10230 points per cluster ) the dimension transport shows quite dramatically and quantitatively the transport of dimension from larger to smaller scale .
we have developed a novel analysis system which is well - suited to the task of precision correlation analysis for general measure distributions and especially for systems which exhibit clustering or other self - similar behavior . by extending the rnyi - entropy concept to a locally - defined function of scale
we are able to establish a more complete picture of data correlations and make precision comparisons among data , simulations and model distributions in the context of information theory .
comparison of monte carlo results and analytic distributions have led to a detailed understanding of scale - local entropy measures .
analysis of simulated clustering data suggests the power of this method in the quantification of scale - dependent correlation structure .
the authors would like to thank all of the people who have contributed to and supported this work .
in particular , we would like to thank dhammika weerasundara who was instrumental in the development of scale - local entropy methods , as well as stephen bailey , justin prosser , and curtis reynolds who helped implement several applications of this analysis .
5 baker g l and gollub j p 1990 _ chaotic dynamics _
( cambridge : cambridge university ) grassberger p 1983 _ phys .
lett . _ a * 97 * 224 - 230 kittel c and kroemer h 1980 _ thermal physics _
( new york : freeman ) pando j and fang l 1998 _ phys .
* 3553 - 3601 pantin e and starck j - l 1996 _ astron . astrophys .
ser . _ * 118 * 575 - 585 reid j g 2002 event - by - event analysis methods and applications to relativistic heavy - ion collision data _ doctoral dissertation _
nucl - ex/0302001 rnyi a 1960 _ mta iii .
* 10 * 251 - 282 trainor t a 1998 scale - local topological measures , _ preprint _ , university of washington , 1998 | a novel method for correlation analysis using scale - dependent rnyi entropies is described .
the method involves calculating the entropy of a data distribution as an explicit function of the scale of a @xmath0-dimensional partition of @xmath0-cubes , which is dithered to remove bias .
analytic expressions for dithered scale - local entropy and dimension for a uniform random point set are derived and compared to monte carlo results .
simulated nontrivial point - set correlations representing condensation and clustering are similarly analyzed .
.2 in keywords : scale , entropy , dimension , information , fractal , complex system , phase transition |
at finite temperature ( @xmath8 ) nuclear matter shows two different phases : a low-@xmath8 hadronic phase , in which chiral symmetry is broken , and a high-@xmath8 phase , in which chiral symmetry is restored and quarks and gluons are unbounded @xcite .
since the @xmath9 and @xmath10 quarks are very light , a great amount of work has been devoted to the study of qcd with @xmath11 light flavors , @xmath12 being the physically interesting case . in this limit
the qcd lagrangian is invariant under @xmath13 and @xmath14 transformations .
since @xmath15_{l , r } , \label{unsun}\ ] ] and the group @xmath16 is isomorphic to the group @xmath17 of vector and axial @xmath18 transformations , the classical symmetry group of the theory can be written as @xmath19_l \otimes [ { \rm su}(n_f)/\mathbb{z}(n_f)]_r .
\label{totsym}\ ] ] the vector subgroup @xmath20 corresponds to the quark - number conservation and it is not expected to play any role at the transition . the @xmath4 symmetry is broken to @xmath21 by quantum fluctuations , since the divergence of the corresponding current presents a quantum anomaly proportional to the topological charge density .
this reduces the relevant symmetry to @xcite @xmath22/\mathbb{z}(n_f)_v .
\label{relsymm}\ ] ] at zero temperature , the hadronic spectrum shows that this symmetry is spontaneously broken to @xmath23 with @xmath24 goldstone particles ( pions and kaons ) and a nonzero quark condensate @xmath25 .
the large mass difference between the pseudoscalar flavor singlet and nonsinglet mesons , such as @xmath26 and @xmath27 , which have the same quark content , reflects the quantum breaking of the @xmath4 symmetry . at finite temperature
, a phase transition occurs at a critical temperature @xmath5 , @xmath28 mev for @xmath12 . above @xmath5 ,
chiral symmetry is restored and the quark condensate vanishes .
therefore , the symmetry - breaking pattern at the chiral transition is expected to be @xmath22/\mathbb{z}(n_f)_v \rightarrow { \rm su}(n_f)_v/\mathbb{z}(n_f)_v , \label{sym1}\ ] ] with a matrix - like order parameter given by the expectation value of the quark bilinear @xmath29 . in the case of two flavors ,
i.e. @xmath12 , the symmetry - breaking pattern ( [ sym1 ] ) is equivalent to that of the o(4 ) vector model , i.e. , to o(4)@xmath30o(3 ) @xcite .
thus , in the case of a continuous transition , the critical behavior of the model with two massless flavors is expected to belong to the three - dimensional ( 3d ) o(4 ) universality class .
the symmetry - breaking pattern at the transition significantly changes if also the @xmath4 symmetry is restored .
the anomaly effects breaking the @xmath4 symmetry are related to the topological properties of qcd .
semiclassical instanton calculations predict a substantial suppression of the instanton density for @xmath31 , where the dilute instanton gas ( dig ) model is expected to provide a reliable approximation @xcite .
for example , in qcd with @xmath11 light flavors of mass @xmath32 , the topological susceptibility @xmath33 is expected to decay asymptotically as @xcite @xmath34 where @xmath35 is the number of colors . for @xmath36 and @xmath12 we have @xmath37 .
although @xmath33 vanishes in the massless limit , the dirac zero modes associated with the instantons induce a residual contribution to the @xmath4 symmetry breaking , giving rise to a difference between the susceptibilities of the so - called @xmath38 and @xmath39 channels at high @xmath8 @xcite , which behaves as @xmath40 in the chiral massless limit .
the breaking of the @xmath4 symmetry at finite @xmath8 , and its role at the chiral transition , has been much investigated .
monte carlo ( mc ) simulations of lattice qcd find a substantial suppression of the @xmath4 anomaly effects at large @xmath8 , as predicted by the dig model .
these results are supported by numerical investigations of pure @xmath41 gauge theories , which show that the topological susceptibility is rapidly suppressed above the deconfinement transition , see , e.g. , ref .
@xcite and references therein , and that the dig regime sets in quite early for @xmath42 @xcite .
there are also some claims of an exact restoration of the @xmath4 symmetry at the chiral transition @xcite .
it is thus worth investigating the nature of the finite-@xmath8 chiral transition in the case the @xmath4 symmetry is effectively restored , and the relevant symmetry - breaking pattern is @xmath43/{\rm u}(1)_v \rightarrow { \rm u}(n_f)_v/{\rm u}(1)_v , \label{sym2}\ ] ] instead of that reported in eq .
( [ sym1 ] ) . up to now
we have discussed the case of the model with @xmath11 massless flavors . however , in nature quarks have a finite mass . since @xmath9 and @xmath10 quarks
are very light , one expects that the correct physical behavior can be obtained by considering their masses as a perturbation in the theory with @xmath12 . according to renormalization - group ( rg ) theory ,
if the transition is continuous in the chiral massless limit , then an analytic crossover is expected for nonzero values of the quark masses @xmath44 , because the quark masses act as external fields coupled to the order parameter .
still , the presence of a close continuous transition gives rise to scaling relations depending on the fermion mass @xmath45 and on the reduced temperature @xmath46 .
for instance , the fermion condensate is expected to scale as @xmath47 where @xmath39 , @xmath48 and @xmath49 are appropriate critical exponents determined by the universality class of the transition , see e.g. refs .
@xcite , and @xmath50 is an universal scaling function ( apart from trivial normalizations ) . on the other hand ,
a first - order transition is generally robust against perturbations .
therefore , if the massless theory undergoes a first - order transition , we expect a first - order transition also for small nonvanishing values of the masses , up to an endpoint @xmath51 , around which a 3d ising critical behavior is expected . for larger fermion masses
the phase transition disappears and we have an analytic crossover as well . to make contact with experiments , it is also necessary to take into account the massive strange quark @xmath52 , whose mass ( @xmath53 mev ) is comparable with @xmath5 . since the transition is expected to be of first order for @xmath54 light degenerate quarks , we also expect a first - order transition when increasing @xmath55 ( keeping @xmath56 ) , at least for sufficiently small values of @xmath55 . for larger values of @xmath55
there are two possibilities , depending on the nature of the transition for @xmath12 degenerate quarks , corresponding to the limit @xmath57 . in one case
we may have a first - order transition line which extends for all values of @xmath55 .
alternatively , the first - order transition line extends up to a finite @xmath58 , then the transition becomes continuous for @xmath59 , and in particular in the limit @xmath60 ; @xmath58 is a tricritical point , separating the first - order transition line from the critical line , which implies that the critical behavior for @xmath61 should be described by mean - field theory , with logarithmic corrections .
the nature of the chiral transition has been extensively studied . in spite of several mc studies of different lattice qcd formulations with two light quarks ,
the nature of the chiral transition is still controversial .
some mc results favor a continuous transition , but are not sufficiently accurate to clearly identify the corresponding universality class .
other mc studies report instead evidence of a first - order transition .
for quark masses close to their physical values , the results of mc simulations support a crossover scenario : the low-@xmath8 and high-@xmath8 regimes are not separated by a phase transition , but rather by a crossover region in which the thermodynamic quantities change rapidly , but continuously , in a relatively narrow temperature interval .
the universal features of the chiral transition can be investigated within the rg framework @xcite .
they are determined by a few global properties , such as the space dimensionality @xmath10 ( @xmath62 for the finite-@xmath8 qcd transition ) , the nature and the symmetry of the order parameter ( a complex matrix related to the bilinear quark operators @xmath63 ) , and the symmetry - breaking pattern ( which is given by eqs .
( [ sym1 ] ) or ( [ sym2 ] ) depending on the role played by the @xmath4 anomaly ) .
for this purpose one considers the rg flow in the space of lagrangians which satisfy the above - reported general properties and determines the fixed points ( fps ) of the flow . in the absence of a stable fp , only first - order transitions between the disordered and ordered phases are possible . on the other hand ,
if a stable fp exists , the transition may be continuous , and the usual critical exponents @xmath64 , @xmath26 , etc ... are related to the eigenvalues of the linearized flow around the fp .
however , it is important to stress that , even in the presence of a stable fp , some systems may still undergo a first - order transition .
from the rg point of view , this occurs in systems which are not in the attraction domain of the stable fp . to determine the rg behavior of the model
, one can use standard perturbative field - theoretical approaches @xcite .
the first rg study of the effective model with symmetry - breaking pattern ( [ sym2 ] ) was presented by pisarski and wilczek @xcite , who performed a one - loop calculation within the @xmath65 expansion , @xmath66 , finding no stable fp close to @xmath67 .
this result suggests a first - order transition for the model with symmetry - breaking pattern ( [ sym2 ] ) .
however , subsequent analyses of the rg flow directly in three dimensions , based on high - order perturbative expansions ( up to six loops ) , have provided the evidence of a stable 3d fp @xcite .
thus , the transition may also be continuous if the @xmath4 symmetry is effectively restored at @xmath5 .
however , its universality class differs from the 3d @xmath68 universality class .
table [ summarytqcd ] summarizes the results of these rg analyses , reporting the possible transitions for various values of @xmath11 , and for the two symmetry breaking patterns ( [ sym1 ] ) and ( [ sym2 ] ) .
when a continuous transition is possible , the corresponding universality class is reported .
.summary of the rg predictions for the finite-@xmath8 qcd transition , as a function of the number @xmath11 of light flavors .
we distinguish two cases , depending whether the u(1)@xmath0 symmetry is broken or effectively restored .
when a continuous transition is possible , we specify the corresponding 3d universality class by reporting its symmetry - breaking pattern . [
cols="^,^,^ " , ] in this paper we extend previous field - theoretical studies @xcite of the 3d @xmath69 universality class .
our purpose is to provide accurate predictions for the critical features of this universality class , for which we know much less compared with the @xmath70 vector universality classes .
we study the rg flow of the multiparameter @xmath3 theory with the same symmetry breaking . for this purpose ,
we consider two different 3d perturbative schemes : the massive zero - momentum ( mzm ) scheme @xcite and the 3d minimal subtraction scheme @xmath71 without @xmath65 expansion @xcite .
the resummation of the perturbative expansions of the @xmath48 functions ( known to six and five loops in the mzm and @xmath71 scheme , respectively ) allows us to compute the rg trajectories starting from the unstable gaussian fp of the free theory .
they approach a stable fp in both schemes for an extended region of bare parameters . moreover
, we estimate the critical exponents by computing the expansions of appropriate rg functions and evaluating them at the stable fp .
the paper is organized as follows . in sec .
[ secrgarg ] we review the universality and rg arguments which we use to investigate the finite-@xmath8 transition in qcd with @xmath12 light flavors . in particular , we define the effective theory that is relevant for the model with symmetry - breaking pattern ( [ sym2 ] ) . in sec .
[ rgflux ] we study the rg flow in the space of the renormalized couplings ; in particular we determine the rg trajectories that start at the unstable gaussian fp of the free quadratic theory and flow towards the stable fp controlling the critical behavior at the transition .
moreover , we determine the critical exponents by evaluating appropriate rg functions at the stable fp . finally , in sec . [ conclusions ] we draw our conclusions .
[ msseries ] we report the perturbative expansions used in the paper to determine the rg flow and the critical exponents .
the nature of the finite-@xmath8 chiral transition in qcd can be investigated using universality and rg arguments @xcite . in this section
we review them , focussing on the finite-@xmath8 transition in qcd with two light flavors .
let us first assume that the phase transition at @xmath5 is continuous for vanishing quark masses . in this case
the length scale of the critical modes diverges approaching @xmath5 , becoming eventually much larger than @xmath72 , which is the size of the euclidean `` temporal '' dimension at @xmath5 .
therefore , the asymptotic critical behavior is associated with a 3d universality class with the same symmetry breaking pattern and the order parameter is an @xmath73 complex - matrix field @xmath74 , related to the bilinear quark operators @xmath63 .
nonvanishing quark masses can be accounted for by an external field coupled to the order parameter . to determine the critical behavior we consider the most general landau - ginzburg - wilson ( lgw ) @xmath3 theory compatible with the given symmetry breaking . if eq .
( [ sym2 ] ) holds , the theory is given by @xmath75 where the field @xmath74 is a generic @xmath73 complex matrix .
the symmetry is u@xmath76u@xmath77 , which breaks to u@xmath78 if @xmath79 , thus providing the lgw theory relevant for qcd with two light flavors .
the reduction of the symmetry to su@xmath76su@xmath77 for qcd , due to the axial anomaly , can be achieved by adding additional quadratic and quartic terms containing the determinant of the field @xmath80
we return to this point later .
the critical behavior at a continuous transition is controlled by the fps of the rg flow , which are determined by the common zeroes of the @xmath48-functions associated with the quartic parameters . to study the rg flow of the @xmath3 theory ( [ lun ] ) , we consider two different perturbative schemes . in the massive zero - momentum ( mzm ) scheme @xcite
one performs the perturbative expansion directly in three dimensions , in the critical region of the disordered phase .
the mzm perturbative expansions of the @xmath48 functions and of the critical exponents have been computed to six loops , requiring the computation of approximately 1000 feynman diagrams .
the six - loop series of the @xmath48 functions were reported in ref .
@xcite , here we also report those of the rg functions associated with the critical exponents , in app . [ sermzm ] . in the 3d @xmath71 scheme one considers the massless critical theory : one uses dimensional regularization and the modified minimal - subtraction prescription , thus the rg functions are obtained from the divergences appearing in the perturbative expansion of the correlation functions @xcite . in the standard @xmath65-expansion scheme @xcite ,
the fps , i.e. , the common zeroes of the @xmath48-functions , are determined perturbatively as expansions in powers of @xmath81 , while exponents are obtained by expanding the corresponding rg functions computed at the fp in powers of @xmath65 .
physical results are then obtained by extrapolating the results to @xmath62 .
this procedure assumes the existence of a fp for @xmath82 , i.e. , close to four dimensions .
therefore , it allows one to determine only those three - dimensional fps which can be defined , by analytic continuation , close to four dimensions .
other fps , which do not have a four - dimensional counterpart , can not be detected .
this problem is overcome by the 3d @xmath83 scheme without @xmath65 expansion @xcite .
the rg functions @xmath84 and @xmath85 are the @xmath83 functions .
however , @xmath86 is no longer considered as a small quantity , but it is set equal to its physical value ( @xmath87 in our case ) before computing the fps .
this provides a well defined 3d perturbative scheme which allows us to compute universal quantities , without the need of expanding around @xmath67 @xcite . in the @xmath83 scheme
the rg series of the @xmath48 functions are known up to five loops @xcite ; here we also present the five - loop series of the rg functions associated with the critical exponents .
they are reported in app .
[ serms ] .
the physically relevant results are obtained by resumming the perturbative expansions ( which are divergent but borel summable ) , using methods that take into account their large - order behavior , which is computed by semiclassical ( hence , intrinsically nonperturbative ) instanton calculations @xcite .
for the model ( [ lun ] ) the large - order behavior is discussed in refs .
the method we use is described in refs .
resummations depend on two parameters , which are optimized in the procedure @xcite .
in this section we study the 3d rg flow of model ( [ lun ] ) in the case relevant for qcd with two light flavors , i.e. for @xmath89 and @xmath79 .
[ zeroesu22 ] provides a sketch of the locations of the stable and unstable fps .
the rg trajectories , starting from the unstable gaussian fp ( denoted by @xmath90 in fig .
[ zeroesu22 ] ) of the quadratic theory , flow toward a nontrivial fp ( denoted by @xmath91 ) for an extended region of quartic bare parameters @xmath92 and @xmath93 , which implies the stability of the fp .
we then determine the critical exponents at the stable fp .
-functions @xmath94 and @xmath95 associated with the quartic couplings of the lagrangian ( [ lun ] ) for @xmath89 .
for @xmath96 the @xmath48-functions have three common zeroes , corresponding to three fps : the gaussian ( g ) and o(8 ) ( o ) fps along the @xmath97 axis are unstable , while the fp ( s ) with @xmath98 and @xmath99 is stable . ]
we first consider the mzm scheme , where one expands in powers of the zero - momentum renormalized quartic couplings .
the theory is renormalized by introducing a set of zero - momentum conditions for the one - particle irreducible two - point and four - point correlation functions of the 2@xmath1002 matrix - like field @xmath101 : @xmath102 , \quad \label{ren1 } \\ & & \gamma^{(4)}_{a_1a_2,b_1b_2,c_1c_2,d_1d_2}(0 ) = 2 \pi z_\phi^{-2 } m^{4-d } \times \label{ren2 } \\ & & \quad \times \left(u u_{a_1a_2,b_1b_2,c_1c_2,d_1d_2 } + v v_{a_1a_2,b_1b_2,c_1c_2,d_1d_2}\right ) , \nonumber \end{aligned}\ ] ] where @xmath103 is the renormalization constant of the order - parameter field @xmath80 , and @xmath104 are appropriate form factors defined so that @xmath105 and @xmath106 at the leading tree order ( more details are reported in ref .
@xcite , where the coupling @xmath107 were denoted by @xmath108 ) .
the fps of the theory are given by the common zeroes of the callan - symanzik @xmath48-functions @xmath109 the resummation of the six - loop series of the @xmath48 functions , as outlined in refs .
@xcite , finds a fp ( point @xmath91 in fig .
[ zeroesu22 ] ) at @xcite @xmath110 beside the unstable gaussian fp at @xmath111 and the o(8 ) fp along the @xmath97 axis , see fig .
[ zeroesu22 ] .
a fp is stable if all eigenvalues of the corresponding stability matrix , @xmath112 ( where @xmath113 corresponds to @xmath107 ) have positive real part .
the numerical analysis of the stability matrix at the fp ( [ mzmfp ] ) favours its stability @xcite . in the following we provide a more direct evidence of the stability of this fp showing that the rg trajectories in the space of the renormalized couplings flow towards this fp for an extended region of bare parameters .
physically , the values @xmath114 and @xmath115 at a stable fp are the two independent ( rg invariant ) couplings which describe the zero - momentum behavior of the quartic correlations in the critical region of the disordered phase @xcite .
the existence of a stable fp is confirmed by the analysis of the @xmath48 functions in the 3d @xmath83 scheme .
the renormalized couplings are again defined from the irreducible four - point correlation function , and the @xmath71 @xmath48 functions are @xmath116 where @xmath117 is the energy scale of this massless scheme .
see ref .
@xcite for more details . in the 3d @xmath118 scheme we set @xmath119 and
then resum the series using borel resummation techniques .
the five - loop series of the @xmath48 functions are reported in app .
[ serms ] .
again a nontrivial common zero of the @xmath48 functions is found at @xcite @xmath120 which is represented by the point @xmath91 in fig .
[ zeroesu22 ] .
note that the renormalized couplings of the mzm and @xmath71 perturbative schemes correspond to different quartic couplings , thus their fp values , cf .
( [ mzmfp ] ) and ( [ msfp ] ) , differ . in order to check the stability of these fps and
determine their attraction domain , we study the rg flow in the space of the renormalized parameters . if @xmath9 and @xmath121 are the renormalized couplings and @xmath93 , @xmath92 the corresponding lagrangian couplings that satisfy @xmath122 and @xmath123 at tree level ( @xmath32 is the zero - momentum mass in the mzm scheme and the renormalization energy scale @xmath117 in the @xmath83 scheme ) , the rg trajectories are determined by solving the differential equations @xmath124 where @xmath125 , with the initial conditions @xmath126 where @xmath52 parametrizes the different rg trajectories in terms of the bare quartic parameters .
note that the initial condition @xmath127 for @xmath121 is required by the theory .
indeed , systems with @xmath128 are associated with transitions with a different symmetry - breaking pattern , i.e. @xmath129 since the stability of the @xmath3 theory ( [ lun ] ) requires @xcite @xmath130 physical systems corresponding to the effective theory ( [ lun ] ) with @xmath79 and @xmath131 are expected to undergo a first - order phase transition . of the bare quartic parameters . ]
scheme , for several values of the ratio @xmath132 of the bare quartic parameters . ]
the rg trajectories for @xmath133 are determined by solving eq .
( [ rgfloweq ] ) after resumming the expansions of the @xmath48 functions . in figs .
[ mzmrgflow ] and [ msrgflow ] we report the rg flow for several values of the ratio @xmath52 , as obtained by a particular choice of the approximants that are used to perform the resummation of the perturbative @xmath48 functions @xcite .
different approximants show analogous qualitative behaviors when they are chosen in the optimal region , defined as outlined in refs .
the rg trajectories in both schemes are attracted by a fp for an extended region of bare quartic parameters @xmath134 , which implies that the fp is stable . in the @xmath83 scheme ,
all trajectories with @xmath135 flow towards the fp given in eq .
( [ mzmfp ] ) , which is indeed the stable fp of the model . in the mzm scheme , we find the same as long as @xmath136 .
the trajectory that corresponds to @xmath137 runs away , into the region in which the perturbative series can no longer be resummed ( the closest borel singularity is on the positive real axis ) , hence we are not able to determine its large-@xmath138 behavior . in any case , both perturbative schemes show the presence of a stable fp .
moreover , both schemes consistently find that the attraction domain of the bare quartic parameters corresponds to @xmath139 .
note that no stable fp is found close to @xmath67 , in agreement with the one - loop @xmath65-expansion calculation of ref .
@xcite , see also ref .
@xcite . however , the extension of this result to the relevant @xmath62 dimension fails .
this is not the only physically interesting case in which @xmath65-expansion calculations fail to provide the correct physical picture in three dimensions .
for example , this also occurs for the ginzburg - landau model of superconductors , in which a complex scalar field couples to a gauge field : although @xmath65-expansion calculations do not find a stable fp @xcite , thus predicting first - order transitions , it is now well established ( see , e.g. , refs .
@xcite ) that 3d systems described by the ginzburg - landau model can also undergo a continuous transition this implies the presence of a stable fp in the 3d ginzburg - landau theory in agreement with experiments @xcite .
other examples are provided by the lgw @xmath3 theories describing frustrated spin models with noncollinear order @xcite and the @xmath140he superfluid transition form the normal to the planar phase @xcite .
finally , we also mention that the rg flow of the u(2)@xmath88u(2 ) scalar theory has been also studied by methods based on approximate solutions of functional rg equations .
@xcite they have not found evidence of a stable fp , but they were limited to approximations keeping only the first terms of the derivative expansion of the effective action .
we now compute the critical exponents by evaluating the corresponding rg functions at the stable fp . in the mzm scheme
they are given by @xmath141 where @xmath103 and @xmath142 are the renormalization functions of the field @xmath80 and of the quadratic operator @xmath143 , respectively @xcite .
the six - loop series are reported in app .
[ sermzm ] .
we have performed an analogous calculation in the @xmath71 scheme .
the perturbative five - loop series are reported in app .
[ serms ] .
the critical exponents are obtained by evaluating the resummed rg functions at the stable fp . in particular , @xmath144^{-1}. \label{exprel}\end{aligned}\ ] ] resumming the perturbative series by using the conformal - borel method @xcite , we obtain @xmath145 , \label{mzmest}\\ & \nu=0.76(10 ) , \quad & \eta=0.11(6),\quad [ { \rm 5\,loop}\;{\overline { \rm ms } } ] . \label{msest}\end{aligned}\ ] ] the errors take into account the uncertainty on the location of the fp , the dependence of the results on the resummation parameters and the stability of the estimates with respect to the number of terms in the series @xcite . estimates ( [ mzmest ] ) and ( [ msest ] ) obtained in the two perturbative schemes are fully consistent .
this agreement provides a nontrivial crosscheck of the accuracy of the analysis of the mzm and 3d @xmath146 perturbative series : the resummation of the perturbative expansions in two different schemes give consistent results for the universal quantities .
this fact may be hardly explained as an artefact of the resummation ; it should instead be considered as a robust evidence of the existence of a stable fp describing 3d continuous transitions characterized by the symmetry - breaking pattern u(2)@xmath88u(2)@xmath30u(2 ) .
in this paper we report a detailed study of the rg flow of the @xmath3 model ( [ lun ] ) for @xmath89 , which is relevant for the finite-@xmath8 chiral transition of two - flavor qcd if the @xmath4 symmetry is restored at the chiral transition . for this purpose
we consider two field - theoretical perturbative schemes : the mzm scheme , defined in the disordered massive phase , and the 3d @xmath71 scheme without @xmath65 expansion , which considers the massless critical theory .
extending previous rg studies @xcite , we verify the existence of a stable fp with @xmath147 , which is the relevant domain for the symmetry - breaking pattern ( [ sym2 ] ) .
we study the rg flow in the space of the renormalized quartic couplings , as obtained by the analysis of the perturbative expansions of the @xmath48 functions , computed to six and five loops in the mzm and 3d @xmath148 scheme , respectively . in both cases
the rg trajectories starting from the unstable gaussian fp flow towards a nontrivial stable fp for an extended region of the bare quartic parameters @xmath134 , i.e. for @xmath149 .
this implies that systems corresponding to an effective lagrangian with @xmath150 undergo a continuous transition .
we also estimate the corresponding critical exponents , obtaining consistent results in the two field - theoretical schemes considered , cf .
( [ mzmest ] ) and ( [ msest ] ) . on the other hand , systems corresponding to @xmath151
are expected to undergo a first - order transition .
the existence of a stable fp with @xmath152 implies that the finite-@xmath8 chiral transition of two - flavor qcd can be continuous also if the @xmath4 symmetry is effectively restored at @xmath5 .
although the critical behavior differs from that expected in the case of a substantial @xmath4 symmetry breaking around @xmath5 , which is the 3d @xmath6 vector universality class , we note that differences are small .
for instance , the critical exponents of the @xmath6 universality class @xcite , @xmath153 , @xmath154 , @xmath155 , @xmath156 , @xmath157 and @xmath158 are close to those we have obtained for model ( [ lun ] ) . in the mzm scheme
we obtain @xmath159 , @xmath160 , @xmath161 , @xmath162 , @xmath163 , and @xmath164 . in the @xmath83 scheme
we obtain instead @xmath165 , @xmath166 , @xmath167 , @xmath168 , @xmath169 , and @xmath170 .
thus , only very accurate estimates of the critical exponents can distinguish the two different critical behaviors .
we stress that the existence of a universality class does not exclude that some systems with the same order parameter and symmetry - breaking pattern undergo a first - order transition .
this occurs when the system is outside the attraction domain of the stable fp , i.e. , when the system at the transition is effectively described by the lagrangian ( [ lun ] ) with quartic parameters @xmath93 and @xmath92 belonging to the large region @xmath171 . the nature ( first - order or continuous ) of the transition is nonuniversal , since it depends on the details of the model and not only on the global features that characterize the universality class .
for example , the model considered in ref .
@xcite , which corresponds to two - flavor lattice qed with the same symmetry breaking @xmath1 , shows a first - order transition .
this result can not be extended to all transitions with the same symmetry breaking , because it is indeed possible that this model corresponds to a run - away rg trajectory , while the case relevant to qcd may belong to the attraction domain of the stable fp , thus undergoing a continuous transition .
-@xmath172 plane for the effective model with symmetry - breaking pattern ( [ sym1 ] ) .
the parameter @xmath172 is proportional to the u(1)@xmath0 symmetry breaking , hence for @xmath173 we reobtain the model with symmetry - breaking pattern ( [ sym2 ] ) . on the left panel ,
the multicritical transition at @xmath173 is continuous , on the right panel it is of first order .
thick black lines indicate first - order transitions .
the endpoints of the first - order transition lines correspond to mean - field transitions with logarithmic corrections . ]
the results for model ( [ lun ] ) are also of interest if chiral symmetry is not exactly restored at @xmath5 , but u(1)@xmath0 breaking effects are small . in this case
we can parametrize the effective lagrangian as @xmath174 , \nonumber\end{aligned}\ ] ] where we added all terms up to dimension four which contain the determinant of the @xmath175 order parameter field and leave a residual su(2)@xmath88su(2 ) symmetry . in the context of two - flavor qcd
, we may assume that @xmath176 where @xmath172 parametrizes the effective breaking of the u(1)@xmath0 symmetry . in the @xmath8-@xmath172 plane ,
the u(2)@xmath177@xmath88u(2)@xmath178 transition point becomes a multicritical point @xcite , as lagrangian ( [ lsu2 ] ) contains two quadratic terms . in fig .
[ pd ] we show two possible phase diagrams , depending on the nature of the transition at @xmath173 . in the first case ( left panel of fig .
[ pd ] ) , the transition is always continuous for @xmath179 along the critical line @xmath180 .
but if @xmath181 is small , we may observe a crossover behavior controlled by the u(2)@xmath88u(2 ) multicritical point at @xmath173 : the free energy should behave as @xmath182 , where @xmath183 , @xmath184 , and @xmath185 @xcite .
in practice , if @xmath172 is small , one might observe two different behaviors depending on the distance of @xmath8 from the critical line @xmath180 .
for @xmath186 not too small , the rg flow is influenced by the u(2)@xmath88u(2 ) multicritical point , hence one would observe an effective critical behavior analogous to that for @xmath173 .
as @xmath8 approaches @xmath180 , this crossover behavior disappears and the o(4 ) behavior is eventually observed . in the other case , shown in the right panel of fig .
[ pd ] , we expect first - order transitions to occur also for @xmath181 small , up to an endpoint @xmath187 where mean - field behavior with logarithmic corrections should be observed ; then the continuous transition for larger @xmath188 ( note that @xmath187 does not need to be small ) is expected to belong to the o(4 ) universality class .
the available numerical mc results for qcd with two light flavors do not yet allow us to distinguish between the above scenarios .
finally , we would like to discuss the possible scenarios for the finite-@xmath8 transitions of the qcd - like theory with a large number @xmath35 of colors , widening the parameter space to get further hints for the relevant @xmath36 case .
we first note that the universality arguments based on the global flavor symmetries do not depend on the number of colors , thus they hold for any @xmath189 ( keeping @xmath11 fixed ) as well , including @xmath190 . at large @xmath35 , keeping the number @xmath11 of flavors fixed , we expect a first - order transition corresponding to the deconfinement transition of pure @xmath191 gauge theories for a large number of colors , thus at @xcite @xmath192 where @xmath193 is the string tension .
the presence of @xmath12 fermion flavors , which contribute to @xmath194 according to standard large-@xmath35 scaling arguments @xcite , can not smooth out this transition whose latent heat is @xmath195 @xcite .
this opens the road to other possible scenarios with respect to the standard three - color qcd .
indeed , the chiral symmetry of the fermions may be restored at the same transition point , or we may have another transition at a larger temperature @xcite , like the case of qcd with adjoint fermions @xcite , where the deconfinement and chiral transitions occur at different temperatures . moreover , since the @xmath4 anomaly is suppressed by @xmath196 in the large-@xmath35 limit @xcite , the @xmath4 symmetry breaking is further suppressed at large @xmath35 , as also shown by the behavior @xmath40 , where the exponent @xmath197 , cf .
( [ chisup ] ) . therefore , with increasing @xmath35 , the effective symmetry breaking at the chiral transition with two light flavors should be better and better described by @xmath198 , rather than @xmath199 .
_ acknowledgements .
_ we thank claudio bonati and massimo delia for useful discussions .
in this appendix we report the perturbative series used in the paper to analyze the rg flow of the u(2)@xmath88u(2 ) @xmath3 theory ( [ lun ] ) and estimate the critical exponents , i.e. the six - loop series of the mzm scheme and the five - loop series of the 3d @xmath71 scheme .
the @xmath48 functions of the mzm perturbative schemes have been already reported to six loops in ref .
@xcite . here
we report the six - loop rg functions defined in eq .
( [ etapt ] ) , which allow us to evaluate the critical exponents through eqs .
( [ exprel ] ) .
they are given by @xmath201 and @xmath202 we report the perturbative series in the @xmath83 scheme .
the @xmath48 functions were computed in ref .
@xcite , where they were explictly reported up to three loops for generic u(@xmath203)@xmath88u(@xmath204 ) models . here
we report the five - loop series of the @xmath48 functions and of the rg functions associated with the critical exponents for the case relevant for qcd with two flavors , i.e. the @xmath3 theory ( [ lun ] ) with @xmath89 .
a. bazavov , t. bhattacharya , m. i. buchoff , m. cheng , n. h. christ , h .-
ding , r. gupta , p. hegde , c. jung , f. karsch , z. lin , r. d. mawhinney , s. mukherjee , p. petreczky , r. a. soltz , p. m. vranas , and h. yin , phys .
d * 86 * , 094503 ( 2012 ) .
a. ali khan , s. aoki , r. burkhalter , s. ejiri , m. fukugita , s. hashimoto , n. ishizuka , y. iwasaki , k. kanaya , t. kaneko , y. kuramashi , t. manke , k. nagai , m. okamoto , m. okawa , a. ukawa , and t. yoshi ( cp - pacs collaboration ) , phys .
d * 63 * , 034502 ( 2000 ) . c. bonati , p. de forcrand , m. delia , o. philipsen , and f. sanfilippo , proc .
lattice2011 ( 2011 ) 189 .
a. bazavov , t. bhattacharya , m. cheng , c. detar , h .-
ding , steven gottlieb , r. gupta , p. hegde , u. m. heller , f. karsch , e. laermann , l. levkova , s. mukherjee , p. petreczky , c. schmidt , r. a. soltz , w. soeldner , r. sugar , d. toussaint , w. unger , and p. vranas , phys .
d * 85 * , 054503 ( 2012 ) .
m. cheng , n. h. christ , s. datta , j. van der heide , c. jung , f. karsch , o. kaczmarek , e. laermann , r. d. mawhinney , c. miao , p. petreczky , k. petrov , c. schmidt , and t. umeda , phys .
d * 74 * , 054507 ( 2006 ) .
a. bazavov , t. bhattacharya , m. cheng , n. h. christ , c. detar , s. ejiri , s. gottlieb , r. gupta , u. m. heller , k. huebner , c. jung , f. karsch , e. laermann , l. levkova , c. miao , r. d. mawhinney , p. petreczky , c. schmidt , r. a. soltz , w. soeldner , r. sugar , d. toussaint , and p. vranas , phys .
d * 80 * , 014504 ( 2009 ) .
s. borsanyi , z. fodor , c. hoelbling , s. d katz , s. krieg , c. ratti , and k. k. szabo ( wuppertal - budapest collaboration ) , j. high energy phys . * 09 * , 073 ( 2010 ) .
f. basile , a. pelissetto , and e. vicari , pos ( lat2005 ) 199 ( 2005 ) .
the approximants we use depend on two parameters @xmath209 and @xmath210 ; we use here the notations of refs . @xcite .
in particular , the rg trajectories reported in figs . [ mzmrgflow ] and [ msrgflow ] are obtained by using @xmath211 and @xmath212 for @xmath94 and @xmath213 and @xmath212 for @xmath95 in both schemes . the first analysis of the mzm @xmath48 functions , reported in ref .
@xcite , was performed in the domain @xmath214 , @xmath215 , and found no evidence of a fixed point beside the o(8 ) fp that occurs in the model with @xmath216 .
subsequent analyses @xcite , however , found an additional stable fixed point outside this region ( point @xmath91 in fig .
[ zeroesu22 ] ) , located at @xmath217 and @xmath218 .
the errors reflect the stability of the results of the different resummations with respect to the parameters @xmath210 and @xmath209 ( see ref .
@xcite ) and the number of perturbative terms considered .
they are obtained by considering how the estimates change when @xmath210 and @xmath209 vary within a reasonably large region around their optimal values ( see ref .
@xcite for a detailed description of the approach ) and when the resummation is applied to the perturbative series with five ( mzm scheme ) or four ( @xmath83 scheme ) terms . | we investigate the nature of the finite - temperature chiral transition in qcd with two light flavors , in the case of an effective suppression of the u(1)@xmath0 symmetry breaking induced by the axial anomaly , which implies the symmetry breaking @xmath1 , instead of @xmath2 . for this purpose
, we perform a high - order field - theoretical perturbative study of the renormalization - group flow of the corresponding three - dimensional multiparameter landau - ginzburg - wilson @xmath3 theory with the same symmetry - breaking pattern .
we confirm the existence of a stable fixed point , and determine its attraction domain in the space of the bare quartic parameters .
therefore , the chiral qcd transition might be continuous also if the @xmath4 symmetry is effectively restored at @xmath5
. however , the corresponding universality class differs from the @xmath6 vector universality class which would describe a continuous transition in the presence of a substantial @xmath4 symmetry breaking at @xmath5 .
we estimate the critical exponents of the @xmath7 universality class by computing and analyzing the corresponding perturbative expansions .
these results are important to discriminate among the different scenarios for the scaling behavior of qcd with two light flavors close to the chiral transition . |
the present day space and velocity distributions of galactic isolated old neutron stars ( ns ) , are the subject of many studies . among the reasons for deriving the positional and kinematical properties of isolated old nss was the suggestion that gamma - ray bursts ( grbs ) may originate from galactic nss ( e.g. , mazets et al .
this was refuted by the homogeneous sky distribution of grbs ( e.g. , meegan et al . 1992 ) , and later on by the discovery of their cosmic origin ( e.g. , metzger et al .
1997 ; van paradijs et al . 1997 ) . another exciting possibility , that re - ignited these efforts , was to detect isolated old nss in soft x - ray radiation that may be emitted as they slowly accrete matter from the interstellar medium ( ism ; ostriker , rees & silk 1970 ; shvartsman 1971 ) .
treves & colpi ( 1991 ) , and blaes & madau ( 1993 ) predicted that @xmath5@xmath6 isolated old ns accreting from the ism would be detected by the _
position sensitive proportional counters ( pspc ) all sky survey ( voges et al .
although intensive searches for these objects were carried out ( e.g. , motch et al .
1997 ; maoz , ofek , & shemi 1997 ; haberl , motch , & pietsch 1998 ; rutledge et al . 2003 ; ageros et al . 2006 ) , only a handful of candidates were found . however , these are presumably young cooling nss , rather than old nss whose luminosity is dominated by accretion from the ism ( e.g. , neuhuser & trmper 1999 ; popov et al .
2000 ; treves et al .
2001 ) . apparently , the reasons for the rareness of these objects in the _ rosat _ source catalog is that their typical velocities were underestimated by earlier studies ( i.e. , narayan & ostriker 1990 ) .
it is also possible that their accretion rate is below the bondi - hoyle rate ( e.g. , colpi et al .
1998 ; perna et al . 2003 ; toropina et al . 2001 , 2003 , 2005
; see however arons & lea 1976,1980 ) , or that these nss are in an ejection stage ( e.g. , colpi et al .
1998 ; livio et al . 1998 ; popov & prokhorov 2000 ) . during the last several years , new classes of radio transients were found ( e.g. , bower et al . 2007 ; matsumura et al . 2007 ; niinuma et al . 2007 ;
kida et al .
2008 ; see also levinson et al . 2002 ; gal - yam et al .
bower et al . (
2007 ) found several examples of these transients in a single small field of view , and showed that they have time scales above 20min but below seven days , and they lack any quiescent x - ray , visible - light , near - infrared ( ir ) , and radio counterparts .
recently , ofek et al . (
2009 ) suggested that these transient events may be associated with galactic isolated old nss .
however , testing this hypothesis requires knowledge of the nss space distribution .
there are two approaches in the literature for calculating the theoretical space and velocity distributions of old nss .
paczynski ( 1990 ) , hartmann , woosley , & epstein ( 1990 ) , blaes & rajagopal ( 1991 ) , blaes & madau ( 1993 ) , and posselt et al .
( 2008 ) carried out monte - carlo simulations of ns orbits .
in such simulations , the positions and velocities of ns at birth are integrated assuming some non - evolving galactic potential .
a second approach is to use some sort of semi - analytic approximation in order to estimate the `` final '' ns space and velocity distributions .
frei , huang , & paczynski ( 1992 ) calculated the final vertical density and velocity distributions of nss , assuming that the gravitational potential is a function only of the height , @xmath7 , above the galactic plane . using the epicyclic approximation , blaes & rajagopal ( 1991 ) developed a technique that allows calculation of the full three dimensional velocity distribution of nss .
however , they showed that this method is not adequate for fast moving objects , which constitute the majority of the ns population .
blaes & madau ( 1993 ) used the thin - disk approximation to calculate the space and velocity distributions of nss . in this prescription , the radial motion of nss are controlled by the galactic potential in the galactic disk , regardless of the vertical height , @xmath7 , of the ns above / below the galactic plane .
finally , another solution was presented by prokhorov & postnov ( 1994 ) , who assume that the ergodic hypothesis is correct ( see discussion in binney & tremaine 1987 ) .
madau & blaes ( 1994 ) noted that all these approaches neglect dynamical heating of nss due to encounters with giant molecular clouds , spiral arms , and stellar `` collisions '' ( e.g. , kamahori & fujimoto 1986 , 1987 ; barbanis & woltjer 1967 ; carlberg & sellwood 1985 ; jenkins & binney 1990 ) .
they crudely estimated the order of magnitude of this effect by applying the force - free diffusion equation to the vertical height , @xmath7 , and ns speed distributions .
they found that , at the solar neighborhood , dynamical heating may decrease the local density of nss by a factor of @xmath8 relative to a non - heated population .
in this paper i present the results of monte - carlo orbital simulations of galactic nss .
these simulations improve upon past efforts ( e.g. , paczynski 1990 ; hartmann , epstein , & woosley 1990 , blaes & rajagopal 1991 ; blaes & madau 1993 ; popov et al . 2005 ; and posselt et al .
2008 ) in several aspects .
first , i use up - to - date space and velocity distributions of nss at birth from arzoumanian et al .
( 2002 ) and faucher - gigure & kaspi ( 2006 ) .
second , i assume a birth rate of nss along the galaxy life time , instead of assuming that all the nss were born about 10 gyr ago .
third , i generate a large sample of simulated nss , which is about one to two orders of magnitude larger than those presented by previous efforts .
the structure of this paper is as follows : in [ sec : modeling ] i present the model ingredients for the monte - carlo simulations . in
[ sec : dynheat ] i discuss dynamical heating and show that it can be neglected .
the monte - carlo simulations are described in [ sec : mc ] , and the catalogue of simulated nss are presented in [ sec : cat ] .
the results of these simulations are presented in [ sec : nsres ] , and in [ sec : disc ] i summarize the results and compare them with some of the previous efforts .
in the following subsections , i present the ingredients of the monte - carlo simulations .
these are : nss birth rate ( [ sec : nsbirthrate ] ) ; space and velocity distributions of nss at birth ( [ sec : nsbirthdist ] ) ; and the galactic potential ( [ sec : galpot ] ) . the stellar population in the milky way is composed of at least two major components : a bulge and a disk .
i therefore , simulate two populations of nss . a galactic disk - born population and a galactic bulge population .
for the galactic - disk nss , i assume a continuous constant formation rate in the past 12gyr .
this assumption is motivated by the analysis of rocha - pinto et al .
( 2000 ) that did not find any major trends in the star formation rate in the disk of our galaxy .
i assume that the galactic - bulge nss were born in a single burst 12 gyr ago ( ferreras , wyse , & silk 2003 ; ballero et al .
2007 ; minniti & zoccali 2008 ) .
as i discuss in this section , the simulations presented here consist of @xmath9 disk - born nss and @xmath10 bulge - born nss .
the actual ratio of disk to bulge nss is unknown and may be significantly different than the one assumed here .
therefore , in the resulting catalogue ( [ sec : cat ] ) i specify the origin of each simulated ns and present the results also for each population separately . the total number of nss in the milky way is constrained by the chemical composition of the galaxy .
specifically , the iron content of the galaxy implies that the total number of core collapse supernovae ( sne ; and therefore nss ) that exploded in the milky way is about @xmath1 ( arnett , schramm , & truran 1989 ) .
however , the star formation rate in the galactic disk was approximately constant during the last @xmath11gyr ( rocha - pinto et al .
2000 ; see also noh & scalo 1990 ) . therefore , with the current ns birth rate of up to one in 30 years ( e.g. , diehl et al . 2006 ) , one expects that there will be @xmath12 disk - born nss in our galaxy .
the predictions based on the sn rate ( @xmath12 ) and the chemical evolution ( @xmath13 ) are therefore inconsistent .
ways around this problem include uncertainty in the sn rate ( e.g. , arnett et al .
1989 ) , and difficulty in estimating the star formation history of our galaxy ( keane & kramer 2008 ) . another viable resolution of this discrepancy is that the galactic bulge had a higher star formation rate at earlier times .
the galactic bulge contains a considerable number of stars that were born , apparently , in a @xmath14gyr - long , burst about 12gyr ago ( ferreras , wyse , & silk 2003 ; ballero et al .
2007 ; minniti & zoccali 2008 ) .
i therefore , assume here that in addition to the disk - born nss there are up to @xmath15 nss that were formed in the galactic bulge about 12gyr ago .
this number was selected such that the total number of nss in the galaxy is @xmath1 .
i note that the galactic bulge contains only about 20% of the galactic stars ( e.g. , klypin et al .
2002 ) . however , ballero et al .
( 2007 ) argued that the initial mass function , @xmath16 , of the bulge is skewed toward high masses , with @xmath17 , the power - law index in the stellar mass function ( @xmath18 ; e.g. , salpeter 1955 ) , being @xmath19 .
assuming that stars with masses exceeding 8m@xmath20 produce nss within a relatively short time after their birth , this suggests that the number of nss born in the galactic bulge is a few times larger than the number of nss born in the galactic disk , per unit stellar mass .
therefore , the assumption that the majority of nss were born in the galactic bulge is in rough agreement with the expected fraction of high mass ( @xmath21m@xmath20 ) stars in the bulge , relative to the disk .
nevertheless , as stated before , the ratio between the numbers of bulge - born to disk - born nss is uncertain .
the simulations presented here cover a wide range of possibilities regarding birth place and time scenarios . following , paczynski ( 1990 )
, i adopt a radial birth probability distribution , @xmath22 , which follows the exponential stellar disk : @xmath23 where @xmath24 is the distance from the galactic center projected on the galactic plane , @xmath25kpc is the exponential disk scale length , and @xmath26 is the normalization given by : @xmath27^{-1 } , \label{expdisknorm}\ ] ] where @xmath28kpc , is the disk truncation radius .
yusifov & kk ( 2004 ) estimated the radial distribution ( measured from the galactic center ) of galactic pulsars , and found it to be more concentrated than predicted by the paczynski ( 1990 ) distribution .
however , i note that the supernova remnant galactic radial distribution is in good agreement with the paczynski ( 1990 ) model . for the ns bulge population
i assume that the space distribution at birth has the same functional form and parameters as the ns disk population , but with @xmath29kpc in eqs .
[ expdisk ] and [ expdisknorm ] . for the initial velocity probability distribution ,
i use two different models .
both models are consistent with the observed velocity distribution of young pulsars .
one is a bimodal velocity distribution composed of two gaussians found by arzoumanian et al .
( 2002 ) , and the second is based on a double - sided exponential unimodal velocity distribution ( faucher - gigure & kaspi 2006 ) . *
bimodal velocity distribution : * the first initial velocity distribution we used is from arzoumanian et al . ( 2002 ) .
it consists of two components . the three - dimensional speed , @xmath30
, probability density of this model is : @xmath31 ) } } \label{birthveldist2}\ ] ] where @xmath32 , @xmath33 and @xmath34 are the fractions of the two ns velocity components , @xmath35kms@xmath36 , and @xmath37kms@xmath36 .
* unimodal velocity distribution : * the second model i use for the velocity distribution of nss at birth was advocated by faucher - gigure & kaspi ( 2006 ) . in their model , each one - dimensional ( projected ) velocity component , @xmath38 , consists of a double - sided exponential probability distribution : @xmath39 where @xmath40kms@xmath36 is the mean velocity in each direction , @xmath41 .
i note that drawing the velocity components of simulated nss directly from eq .
[ birthveldist1 ] will result in an aspherical distribution . in order to avoid this problem
the corresponding three - dimensional distribution is required .
faucher - gigure & kaspi ( 2006 ) noted that this distribution is difficult to derive analytically , and did not provide its functional form .
however , it is possible to show that the corresponding spherically symmetric three - dimensional distribution corresponding to eq .
[ birthveldist1 ] is given by @xmath42 in figure [ fig : initveldist_compare ] i show the one - dimensional ( _ dashed _ lines ) and three - dimensional ( _ solid _ lines ) initial velocity probability distributions based on faucher - gigure & kaspi ( 2006 ) velocity distribution ( _ thin black _ lines ) and the arzoumanian et al .
( 2002 ) distribution ( _ thick gray _ lines ) .
several other fits for the pulsars velocity distribution at birth are available ( e.g. , hansen & phinney 1997 ; cordes & chernoff ( 1998 ) ; hobbs et al . 2005 ; zou et al . 2005 ) . however , the arzoumanian et al .
( 2002 ) and faucher - gigure & kaspi ( 2006 ) velocity distributions cover a wide range of possibilities , that reflect the current uncertainties in the pulsars velocity distribution at birth .
the two models are considerably different from each other , mainly in the low and high velocity tails . relative to the arzoumanian et al .
( 2002 ) velocity distribution the faucher - gigure & kaspi ( 2006 ) distribution has excess of nss at low velocities . since the slowest nss are the most efficient accretors from the ism ( e.g. , ostriker et al .
1970 ) , these differences may have a significant impact on the detectability of isolated old nss in x - ray wavebands ( e.g. , blaes & madau 1993 ) .
, width=321 ] the velocity distributions , in eqs .
[ birthveldist2 ] and [ birthveldist1_3 ] , are given relative to a rotating disk .
therefore , i added the disk rotation velocity to these birth velocities @xmath43 @xmath44 @xmath45 where @xmath38 is the velocity vector relative to the rotating disk ( i.e. , the one obtained from eqs .
[ birthveldist2 ] or [ birthveldist1_3 ] ) , @xmath46 is the velocity vector in a non - rotating ( inertial ) reference frame , @xmath47 , @xmath48 is azimuthal angle on the galactic disk ( @xmath49 $ ] ) , which was selected from a uniform random distribution , and @xmath17 , @xmath50 , and @xmath7 are the position in a cartesian , non - rotating , coordinate system , whose origin is the galactic center , and @xmath7 is vertical to the galactic plane . finally , @xmath51 was obtained from @xmath52 where @xmath53 is the galactic potential ( see [ sec : galpot ] ) , at the point of interest . finally , i assume that the vertical height above the galactic mid - plane of nss at birth , @xmath7 , is drawn from a gaussian distribution ( @xmath54 $ ] ) , with @xmath55kpc and @xmath56kpc , for the arzoumanian et al .
( 2002 ) and faucher - gigure & kaspi ( 2006 ) initial velocity distributions , respectively . following paczynski ( 1990 )
, i assume that the galactic potential is composed of a disk component , an spheroidal component , and an halo component .
i also added a central black hole component .
the disk and spheroid components are described by the following potential proposed by miyamoto & nagai ( 1975 ) , @xmath57^{2 } \big\}^{1/2 } } , \label{potential_disk_sphere}\ ] ] where @xmath58 , @xmath59 and @xmath60 are given in equations [ eq : potential_d_par ] and [ eq : potential_s_par ] , and where @xmath61 corresponds to the disk component and @xmath62 corresponds to the spheroid component , and @xmath63 is the gravitational constant .
next , the halo potential is given by @xmath64 , \label{potential_halo}\ ] ] were @xmath65 , @xmath66 , and @xmath67 are listed in eq .
[ potential_h_par ] .
furthermore , i added a component representing the galactic central massive black - hole ( e.g. , ghez et al . 1998 ) : @xmath68 where @xmath69 is given in eq .
[ potential_bh_par ] ( eisenhauer et al . 2005 ) .
the choice of parameters listed here reproduces the observed galactic rotation , local density , and local column density ( see paczynski 1990 for details ) : @xmath70 @xmath71 @xmath72 @xmath73 i note that the black hole contributes about @xmath74 ( @xmath75 ) of the gravitational potential at distance of 1pc ( 10pc ) from the black hole .
therefore , its influence on the galactic potential is negligible and it does not change the fitted parameters in eqs .
[ eq : potential_d_par]-[potential_bh_par ] . however , it may heat nss passing nearby .
finally , the galactic potential is the sum of these four components : @xmath76
dynamical heating , presumably by giant molecular clouds ( e.g. , kamahori & fujimoto 1986 ; 1987 ) , spiral structure ( e.g. , barbanis & woltjer 1967 ; carlberg & sellwood 1985 ; jenkins & binney 1990 ) , and stars , tends to broaden the velocity and spatial distributions of galactic stars ( e.g. , wielen 1977 ; nordstrom et al .
2004 ) . in order to roughly estimate the effect of dynamical heating on nss , madau & blaes ( 1994 ) applied the force - free diffusion equation to the velocity and vertical distance distributions of nss .
they adopted the total diffusion coefficient , @xmath77 ( @xmath78 ) , measured by wielen ( 1977 ; @xmath79km@xmath80s@xmath81gyr@xmath36 ) . in this
, @xmath82 corresponds to the radial direction , and is positive in the direction of the galactic center ; @xmath83 points in the direction of circular rotation ; and @xmath84 is directed towards the north galactic pole . applying heating
, they found that the local density of nss is smaller by about 30% relative to the case of no heating . in the following we discuss this approach in light of new observational data available .
we estimate the diffusion coefficient , @xmath77 , using modern data .
nordstrom et al .
( 2004 ) , estimated the dynamical heating by the galactic disk and showed that it does not saturate after some time as suggested by measurements based on smaller samples ( e.g. , quillen & garnett 2001 ; see also aumer & binney 2009 ) .
approximating dynamical heating by a random walk process , i fit the nordstrom et al .
( 2004 ) measurements with the function ( e.g. , wielen 1977 ) @xmath85 where @xmath86 is the velocity dispersion component at time , @xmath87 , since birth , @xmath88 is the initial velocity dispersion component and @xmath89 .
the data and the best fit curves are shown in figure [ fig : sigmat ] , and the best fit parameters are listed in table [ tab : diffuspar ] . ) along the @xmath82 ( radial ; _ circles _ ) , @xmath83 ( tangential ; _ triangles _ ) and @xmath84 ( vertical ; _ squares _ ) galactic directions as a function of age since birth , @xmath87 , as measured by nordstrom et al .
the lines represent the best fit of equation [ vel_diffus ] to the data .
the best fit parameters are listed in table [ tab : diffuspar ] .
[ fig : sigmat],width=321 ] lccc u & @xmath90 & @xmath91 & @xmath92 + v & @xmath93 & @xmath94 & @xmath95 + w & @xmath96 & @xmath97 & @xmath98 [ tab : diffuspar ] the total diffusion coefficient that i find , @xmath99km@xmath80s@xmath81gyr@xmath36 , is about half of the value found by wielen ( 1977 ) , and used by madau & blaes ( 1994 ) .
next , the initial velocities estimated by narayan & ostriker ( 1990 ) , that was used by madau & blaes ( 1994 ) , are considerably lower than the more recent estimates ( e.g. , arzoumanian et al .
2002 ) . as suggested by eq .
[ vel_diffus ] , diffusion affects mostly low velocity objects .
therefore , dynamical heating is less important than estimated by madau & blaes ( 1994 ) .
furthermore , the approach taken by madau & blaes ( 1994 ) assumes that nss are being affected by diffusion at all times .
however , the scatterers are restricted to the galactic plane .
nss are born with high velocities and spend @xmath100 to @xmath101 of the time at distances larger than 100pc from the galactic plane ( based on the results in [ sec : nsres ] ) .
therefore , they are less susceptible to dynamical heating than disk stars . based on these three arguments ,
i conclude that the importance of dynamical heating was over estimated by madau & blaes ( 1994 ) by at least an order of a magnitude .
nevertheless , dynamical heating may effect some of the slow moving objects , but they are a minority among the galactic nss .
[ vel_diffus ] roughly suggests that dynamical heating is important for nss with speeds smaller than about @xmath102 .
even for @xmath103gyr this gives 60kms@xmath36 .
however , only @xmath104 and @xmath105 of the nss in the arzoumanian et al . ( 2002 ) and faucher - gigure & kaspi ( 2006 ) , initial velocity distributions , respectively , have speeds smaller than 60kms@xmath36 at birth ( relative to their local standard of rest ) .
to solve for ns orbits i integrate the equations of motion @xmath106 using a livermore ordinary differential equation solver ( hindmarsh 1983 ) . the integration is performed in a non - rotating , cartesian coordinates system , the origin of which is the galactic center , and the @xmath7 axis is perpendicular to the galactic plane . for each initial velocity distribution ( i.e. , bimodal or unimodal ; see [ sec : nsbirthdist ] ) , i simulated @xmath0 ns orbits . in each simulation
i randomly drew the nss birth times , positions , and velocities , from the probability distributions described in [ sec : nsbirthrate ] , and [ sec : nsbirthdist ] . as explained before , @xmath9 of these nss
are disk - born , and the rest are bulge - born . at the end of each simulation
i checked if the integration conserved the total energy . in cases in which the energy was not conserved to within @xmath107 , i reran the integration using the same initial conditions , with a refined integration tolerance .
in the second ( and final ) iteration , the energy was conserved to better than @xmath104 in all cases .
the catalogue of initial ( i.e. , at birth ) and final ( i.e. , current epoch ) simulated nss space and velocity components are listed in table [ tab : mc_v2 g ] for the bimodal initial velocity distribution of arzoumanian et al .
( 2002 ) , and in table [ tab : mc_v1e ] for the unimodal initial velocity of faucher - gigure & kaspi ( 2006 ) .
the first column in each table indicates whether the simulated ns belongs to the bulge population ( code 0 ) or disk population ( code 1 ) .
crrrrrrrr|rrrrrr 0 & 12.0&@xmath108 & @xmath109 & @xmath110 & @xmath111 & @xmath112 & @xmath113&@xmath114&@xmath115&@xmath116&@xmath117&@xmath118&@xmath119&@xmath120 + 0 & 12.0&@xmath121 & @xmath122 & @xmath123 & @xmath124 & @xmath125 & @xmath126&@xmath127&@xmath128&@xmath129&@xmath130&@xmath131&@xmath132&@xmath133 + 0 & 12.0&@xmath134 & @xmath122 & @xmath135 & @xmath136 & @xmath137 & @xmath138&@xmath139&@xmath140&@xmath141&@xmath142&@xmath143&@xmath144&@xmath145 + 0 & 12.0&@xmath146 & @xmath147 & @xmath148 & @xmath149 & @xmath150 & @xmath151&@xmath152&@xmath153&@xmath154&@xmath155&@xmath156&@xmath157&@xmath158 + 0 & 12.0&@xmath159 & @xmath160 & @xmath161 & @xmath162 & @xmath163 & @xmath164&@xmath165&@xmath166&@xmath167&@xmath168&@xmath169&@xmath170&@xmath171 [ tab : mc_v2 g ] crrrrrrrr|rrrrrr 1 & 2.3&@xmath172&@xmath173&@xmath174&@xmath175&@xmath176&@xmath177&@xmath178&@xmath179&@xmath180&@xmath181&@xmath182&@xmath183&@xmath184 + 0 & 12.0&@xmath185&@xmath186&@xmath187&@xmath188&@xmath189&@xmath190&@xmath191&@xmath192&@xmath193&@xmath194&@xmath195&@xmath196&@xmath197 + 1 & 2.9&@xmath198&@xmath199&@xmath174&@xmath200&@xmath201&@xmath202&@xmath203&@xmath204&@xmath205&@xmath206&@xmath207&@xmath208&@xmath209 + 1 & 0.6&@xmath210&@xmath211&@xmath212&@xmath213&@xmath214&@xmath215&@xmath216&@xmath217&@xmath218&@xmath219&@xmath220&@xmath221&@xmath222 + 0 & 12.0&@xmath223&@xmath224&@xmath212&@xmath225&@xmath226&@xmath227&@xmath228&@xmath229&@xmath230&@xmath231&@xmath232&@xmath233&@xmath234 [ tab : mc_v1e ] in addition , catalogue of simulated ns positions , distances , radial velocities , and proper motions for an observer located at the `` solar circle '' , moving around the galactic center with a velocity of 220kms@xmath36 is given in tables [ tab : suncentric2 ] and [ tab : suncentric1 ] .
the solar circle is defined to be on the galactic plane ( i.e. , @xmath235kpc ) at the distance of the sun from the galactic center ( @xmath236kpc ; ghez et al .
the catalogue in tables [ tab : suncentric2 ] and [ tab : suncentric1 ] are based on the initial velocity distribution of arzoumanian et al .
( 2002 ) and the faucher - gigure & kaspi ( 2006 ) , respectively .
i note that the radial velocity and proper motions are calculated for a static observer with respect to the local standard of rest ( lsr ; i.e. , the solar motion with respect to the lsr is neglected ) . in order to reduce poisson errors in the local properties of nss ( i.e. , density ; sky surface density ) , tables [ tab : suncentric2 ] and [ tab : suncentric1 ] were produced by calculating the positions of the @xmath0 simulated nss in tables [ tab : mc_v2 g ] and [ tab : mc_v1e ] , respectively , from 100 random locations on the solar circle .
hence , tables [ tab : suncentric2 ] and [ tab : suncentric1 ] list @xmath237 simulated nss .
crrrrrrr 0 & 12.0000 & @xmath238&@xmath239&@xmath240&@xmath241&@xmath242&@xmath243 + 0 & 12.0000 & @xmath244&@xmath245&@xmath246&@xmath247&@xmath248&@xmath249 + 0 & 12.0000 & @xmath250&@xmath251&@xmath252&@xmath253&@xmath254&@xmath255 + 0 & 12.0000 & @xmath256&@xmath257&@xmath258&@xmath259&@xmath260&@xmath261 + 0 & 12.0000 & @xmath262&@xmath263&@xmath264&@xmath265&@xmath266&@xmath267 [ tab : suncentric2 ] crrrrrrr 1 & 2.2753 & @xmath268&@xmath269&@xmath270&@xmath271&@xmath272&@xmath273 + 0 & 12.0000 & @xmath274&@xmath275&@xmath276&@xmath277&@xmath278&@xmath279 + 1 & 2.8584 & @xmath280&@xmath281&@xmath282&@xmath283&@xmath284&@xmath285 + 1 & 0.6194 & @xmath286&@xmath287&@xmath288&@xmath289&@xmath290&@xmath291 + 0 & 12.0000 & @xmath292&@xmath293&@xmath294&@xmath295&@xmath296&@xmath297 [ tab : suncentric1 ]
in this section i present the space and velocity distributions of simulated nss , at the current epoch . in [ sec : genprop ] i discuss the overall distribution of nss in the galaxy , while in [ sec : obsprop ] i discuss their statistical properties as observed from the lsr .
additional specific statistical properties of these objects are discussed in ofek et al .
( 2009 ) in the context of the long - duration radio transients ( bower et al . 2007 ;
kida et al .
the results presented in this section assume there are @xmath298 nss in the galaxy , where @xmath299 is the number of nss in the galaxy , of which @xmath9 were born in the disk and @xmath10 in the bulge .
nss are born with large space velocities , which are typically of the order of the escape velocity from the galaxy . these are presumably the result of kick velocities due to asymmetric supernovae explosions ( e.g. , blaauw 1961 ; lai et al .
therefore , it is expected that a large fraction of the galactic nss will be unbounded to the milky way gravitational potential and some may be found at very large distances . in figure
[ fig : ns_energydist ] , i show the distribution of the total ( kinetic @xmath300 potential ) energy of the simulated nss : @xmath301 , where the nss mass was set to @xmath302m@xmath20 .
panel ( a ) shows the energy distribution based on the simulations using the unimodal initial velocity distribution of faucher - gigure & kaspi ( 2006 ) , while panel ( b ) is for the bimodal distribution of arzoumanian et al .
( 2002 ) . in each panel , the _ thick black solid _ line represents the entire ns population , the _ thin solid _ line shows the disk - born nss , and the _ dashed _ line is for bulge - born nss . .
the different lines shows the distribution of all the nss ( _ solid thick black _ lines ) ; only disk - born nss ( _ solid thin _ lines ) ; and only bulge - born nss ( _ dashed _ lines ) .
[ fig : ns_energydist],width=321 ] using the approximation that all the nss with negative energies are gravitationally bounded to the galaxy ( neglecting heating ) , in table [ tab : unboundedns ] i give the fractions of nss bounded to the galactic gravitational potential .
lccc a2002 & 0.38 & 0.16 & 0.52 + fk2006 & 0.30 & 0.13 & 0.41 [ tab : unboundedns ] given the large fraction of gravitationally - unbounded nss , it is expected that some nss may be found at very large distances , @xmath303 , from the galactic center .
i find that about @xmath304 and @xmath305 of the nss born in the galaxy are currently at distances larger than 1mpc from the galactic center , for the unimodal and bimodal initial velocity distributions , respectively .
for @xmath306mpc@xmath307mpc , i find that the density of nss as a function of @xmath303 is about : @xmath308pc@xmath3 and @xmath309pc@xmath3 , for the bimodal and unimodal initial velocity distributions , respectively . finally , i find that some milky way born nss may be at distances as large as 30 to 40mpc from the galaxy .
i note that the local density , in our galaxy , of nss born in other galaxies , is of the order of @xmath310pc@xmath3 .
this was estimated by calculating the density in the the milky way , of nss born in each galaxy found within 10mpc .
for this , i used a version of the tully ( 1988 ) nearby galaxy catalog ( ofek 2007 ) where the total number of nss in each galaxy was normalized by its total @xmath311-band magnitude , relative to milky way . in figures [ fig : ns_radialdensity ] and [ fig : ns_radialsurfacedensity ] i show the density of nss on the galactic plane , and the surface density of nss projected on the galactic plane , respectively , as a function of distance from the galactic center .
the notations are the same as in figure [ fig : ns_energydist ] .
in addition , in fig .
[ fig : ns_radialsurfacedensity ] , the _ gray solid _ line represents the initial surface distribution of all nss .
furthermore , the vertical _ dashed _ lines mark the distance of the sun from the galactic center , @xmath236kpc ( ghez et al .
as noted before , i refer to this distance from the galactic center , when located on the galactic plane , as the solar circle .
the densities in the solar circle and galactic center are listed in the figure captions .
i note that the fluctuations seen in small and large radii are due to poisson noise . for details regarding line types .
the density is calculated by counting the nss within 100pc from the galactic plane , dividing it by the appropriate volume in each bin , and assuming that the total number of nss in the galaxy is @xmath1 . for the initial velocity distribution of faucher - gigure & kaspi ( 2006 ; panel a ) , the nss density at the solar circle ( distance of 8.0kpc from the galactic center ; ghez et al .
2008 ) is @xmath312pc@xmath3 , of which @xmath10 are disk born and @xmath9 are bulge born . the nss density at the galactic center is : @xmath313pc@xmath3 , of which @xmath314 are disk born and @xmath315 are bulge born . for the initial velocity distribution of arzoumanian et al .
( 2002 ; panel b ) , the density at the solar circle is @xmath316pc@xmath3 , of which @xmath317 are disk born and @xmath318 are bulge born .
the density at the galactic center is : @xmath319pc@xmath3 , of which @xmath320 are disk born and @xmath321 are bulge born .
i note that the fluctuations seen in small and large radii are due to poisson noise .
[ fig : ns_radialdensity],width=321 ] for details .
in addition the _ gray _ lines show the initial surface density distributions . for the initial velocity distribution of faucher - gigure & kaspi ( 2006 ; panel a ) , the nss surface density at the solar circle is @xmath322pc@xmath81 of which @xmath323 are disk born nss , and @xmath324 are bulge - born nss .
the surface density at the galactic center is @xmath325pc@xmath81 of which about @xmath326 are disk born nss , and the rest are bulge - born nss . for the initial velocity distribution of arzoumanian et al .
( 2002 ; panel b ) , the nss density at the solar circle is @xmath327pc@xmath81 of which @xmath328 are disk born nss , and @xmath329 are bulge - born nss .
the corresponding surface density at the galactic center is @xmath330pc@xmath81 of which @xmath331 are disk born nss , and the rest are bulge - born nss .
as before , the plotted densities assume the total number of nss is @xmath1 .
[ fig : ns_radialsurfacedensity],width=321 ] in figure [ fig : ns_gal_zdistsolarcirc ] i show the nss density as a function of height above or below the galactic plane , as measured at the solar circle .
this was calculated by counting the number of simulated nss with distance from the galactic center , projected on the galactic plane , of @xmath332kpc .
the line scheme is the same as in figures [ fig : ns_energydist ] . at the solar circle the scale height .
] of nss is about @xmath333kpc and @xmath334kpc for the arzoumanian et al .
( 2002 ) and faucher - gigure & kaspi ( 2006 ) initial velocity distributions , respectively . , above / below the galactic plane as measured on the solar circle at the current epoch .
line types are like those in fig .
[ fig : ns_energydist ] . for the initial velocity distribution of faucher - gigure & kaspi ( 2006 ; panel a ) , @xmath335 ( @xmath101 ) of the nss are found within 0.9kpc ( 5.6kpc ) from the galactic plane .
for the initial velocity distribution of arzoumanian et al .
( 2002 ; panel b ) , @xmath335 ( @xmath101 ) of the nss are found within 0.7kpc ( 4.4kpc ) from the galactic plane .
[ fig : ns_gal_zdistsolarcirc],width=321 ] shown in fig .
[ fig : ns_radialdensityz ] are the space densities of nss as a function of projected ( on the galactic plane ) distance from the galactic center , for several different galactic height : 0kpc ; 1kpc ; 3kpc ; and 10kpc .
the densities are calculated in slices , parallel to the galactic plane , with semi - width of 0.1kpc . , and as measured at the current epoch .
panel ( a ) shows the densities calculated using the initial velocity distribution of faucher - gigure & kaspi ( 2006 ) . for this model , at the solar circle , the densities are about @xmath336pc@xmath3 , @xmath337pc@xmath3 , @xmath338pc@xmath3 , and @xmath339pc@xmath3 , at galactic heights of 0kpc , 1kpc , 3kpc , and 10kpc , respectively .
panel ( b ) shows the densities calculated using the initial velocity distribution of arzoumanian et al .
( 2002 ) . for this model , at the solar circle , the densities are about @xmath340pc@xmath3 , @xmath341pc@xmath3 , @xmath342pc@xmath3 , and @xmath343pc@xmath3 , at galactic heights of 0kpc , 1kpc , 3kpc , and 10kpc , respectively .
[ fig : ns_radialdensityz],width=321 ] finally , in figure [ fig : ns_speeddist ] i show the initial and final speed distributions as measured relative to an inertial reference frame ( contrary to fig .
[ fig : initveldist_compare ] which shows the initial speed distribution relative to a rotating reference frame ) .
the probabilities in this figure are shown per 1kms@xmath36 bins .
the line scheme is again like the one used in fig .
[ fig : ns_radialsurfacedensity ] . as expected , the typical speeds of nss decrease with time as they , on average , increase their distances from the galaxy and lose kinetic energy . for the unimodal initial velocity distribution , at the current epoch
i estimate that about @xmath344 , @xmath345 , and @xmath346 of the nss have speeds below 100 , 200 , and 1000kms@xmath36 , respectively . for the bimodal initial velocity distribution , at the current epoch
i estimate that about @xmath347 , @xmath329 , and @xmath348 of the nss have speeds below 100 , 200 , and 1000kms@xmath36 , respectively . .
the probabilities are calculated for 1kms@xmath36 bins .
panel ( a ) shows the results for the unimodal initial velocity distribution , while panel ( b ) for the bimodal initial velocity distribution .
[ fig : ns_speeddist],width=321 ] in this subsection i discuss : ( i ) the expectancy distance of the nearest ns to the sun ; ( ii ) the number of young nss in the solar neighborhood ; ( iii ) the proper motion distribution of nearby nss ; and ( iv ) the all - sky distribution of galactic nss . the probability to find a ns within distance @xmath349 from the sun
is given by @xmath350 given the local density of nss , @xmath351 , that i have found in
[ sec : genprop ] , this implies that the expectancy distance of the nearest ns is about @xmath352pc and @xmath353pc for the arzoumanian et al .
( 2002 ) and faucher - gigure & kaspi ( 2006 ) distributions , respectively . for the bimodal initial velocity distribution ( table [ tab : suncentric2 ] ) , within 1kpc from the sun , 63% of the nss are disk - born , and there are about @xmath354 ( @xmath355 ) nss younger than 1myr ( 10myr ) .
on the other hand , for the unimodal initial velocity distribution ( table [ tab : suncentric1 ] ) , within 1kpc from the sun , 57% of the nss are disk - born , and there are about @xmath356 ( @xmath357 ) nss younger than 1myr ( 10myr ) . in figure
[ fig : ns_pm_dist ] , i show the median total proper motion ( _ solid _ lines ) of simulated nss as a function of their distance from an observer located on the solar circle .
the _ dotted _ lines show the lower and upper @xmath358-percentiles of the proper motion distributions .
the _ black _ lines represent the unimodal initial velocity distribution and the _ gray _ lines are for the bimodal initial velocity distribution .
in addition , the _ dots _ show the observed proper motions of known pulsars ( manchester et al . 2005 ) .
-percentiles of the proper motion distributions .
the _ black _ lines represent the unimodal initial velocity distribution and the _ gray _ lines are for the bimodal initial velocity distribution .
in addition , the _ dots _ show the observed proper motion of known pulsars .
[ fig : ns_pm_dist],width=321 ] i note that the largest total proper motion in table [ tab : suncentric2 ] ( i.e. , bimodal initial velocity distribution ) is @xmath359yr@xmath36 , and that @xmath360 nss are expected to have proper motion in excess of @xmath361yr@xmath36 . in table
[ tab : suncentric1 ] , the largest proper motion is @xmath362yr@xmath36 , and about @xmath363 nss are expected to have proper motion in excess of @xmath361yr@xmath36 . in fig .
[ fig : ns_skydistribution]a f i show the sky surface density of nss at the current epoch , for an observer located at the solar circle .
panels ( a)(c ) are for the unimodal initial velocity distribution , while panels ( d)(f ) are for the bimodal initial velocity distribution .
panels ( a ) and ( d ) show the distribution for all the nss ( i.e. , disk- and bulge - born populations ) , panels ( b ) and ( e ) for the disk - born population , and panels ( c ) and ( f ) for the bulge - born population . the maps are shown in the aitoff equal - area projection and galactic coordinate system , where the galactic center is at the center of each map .
the surface densities are normalized assuming that there are @xmath364 disk - born nss , and @xmath15 bulge - born nss . at the positions with the lowest surface density of nss , the poisson errors due to the limited statistics are smaller than about @xmath365 .
i find that the minimum surface density is attained at the direction of the galactic poles , and is about @xmath366deg@xmath81 and @xmath367deg@xmath81 for the unimodal and bimodal initial velocity distributions , respectively .
the maximum surface density is at the direction of the galactic center and it is about @xmath368deg@xmath81 and @xmath369deg@xmath81 , for the unimodal and bimodal initial velocity distributions , respectively .
i note that the differences between the sky surface densities resulting from the two initial velocity distributions , are as large as about @xmath370 .
the milky way s nss content , and in particular the ns space and velocity distributions are important for searching nearby isolated old nss , and studying any ongoing activity from such objects . as a tool for such studies ,
i present a mock catalogue of simulated isolated old nss spatial positions and velocities , at the current epoch .
the catalogue were constructed by integrating the equations of motion of simulated nss in the galactic potential .
the simulations include two populations of nss , one in which the nss were born in the galactic bulge about 12gyr ago , and the second population in which nss are being born at the galactic disk at a constant rate , starting 12gyr ago .
the combined ns population assumes that @xmath10 of the nss originated in the bulge and the rest in the disk .
although we do not know what is the exact number of galactic nss , and what was their position - dependent birth rate , these two populations provide a wide range of initial conditions .
i generated two catalogue of simulated nss .
each catalog contains @xmath0 objects .
the two catalogue utilize different initial velocity distributions of nss .
one catalog ( table [ tab : mc_v2 g ] ) uses the initial velocity distribution of arzoumanian et al .
( 2002 ) , while the other ( table [ tab : mc_v1e ] ) uses the initial velocity distribution of faucher - gigure & kaspi ( 2006 ) .
also derived are catalogue of simulated ns positions and proper motions with respect to an observer found at the solar circle ( tables [ tab : suncentric2 ] and [ tab : suncentric1 ] ) .
the space distribution at the current epoch obtained by the two initial velocity distributions implemented here , are somewhat different .
for example , i find that the resulting sky surface density , based on the two different initial velocity distributions , differs by up to @xmath370 .
the main differences between the two velocity distributions are in the low- and high - tails of the nss velocity distribution ( see fig . [
fig : initveldist_compare ] ) .
lccl @xmath371 & 0.42 & 240 & this work ( disk@xmath300bulge ) a2002 + @xmath372 & 0.20 & 200 & this work ( disk@xmath300bulge ) fk2006 + @xmath373 & 0.20 & & paczynski ( 1990 ) + @xmath374 & 0.27 & 43 & blaes & madau ( 1993 ) + @xmath375 & 0.50 & 69 & madau & blaes ( 1994 ) + @xmath376 & & 140 & perna et al . ( 2003 ) + @xmath377 & 0.4 & & popov et al . (
2005 ) [ tab : comp ] the space and velocity distributions of galactic nss were estimated by several past works . in table
[ tab : comp ] , i compare some of the basic results of the simulations presented here ( e.g. , local density of nss ; scale height ) , with the ones obtained by previous efforts .
i thank reem sari , ehud nakar , orly gnat , avishay gal - yam , sterl phinney , and mansi kasliwal for valuable discussions .
i also thank an anonymous referee for valuable comments .
support for program number hst - go-11104.01-a was provided by nasa through a grant from the space telescope science institute , which is operated by the association of universities for research in astronomy , incorporated , under nasa contract nas5 - 26555 .
hindmarsh , a. c. , 1983 , odepack , a systematized collection of ode solvers , in scientific computing , r. s. stepleman et al .
( eds . ) , north - holland , amsterdam , 1983 ( vol . 1 of imacs transactions of scientific computation ) , pp . | i present the results of monte - carlo orbital simulations of galactic neutron stars ( nss ) .
the simulations take into account the up - to - date observed ns space and velocity distributions at birth , and account for their formation rate .
i simulate two populations of nss .
objects in the first population were born in the galactic disk at a constant rate , in the past 12gyr .
those in the second population were formed simultaneously 12gyr ago in the galactic bulge .
i assume that the nss born in the galactic disk comprise 40% of the total ns population .
since the initial velocity distribution of nss is not well known , i run two sets of simulations , each containing @xmath0 simulated nss .
one set utilizes a bimodal initial velocity distribution and the other a unimodal initial velocity distribution , both are advocated based on pulsars observations . in light of recent observational results , i discuss the effect of dynamical heating by galactic structure on ns space and velocity distributions and show it can be neglected .
i present catalogue of simulated ns space and velocity vectors in the current epoch , and catalogue of positions , distances and proper motions of simulated nss , relative to the sun .
assuming there are @xmath1 nss in the galaxy , i find that in the solar neighborhood the density of nss is about @xmath2pc@xmath3 , and their scale height is about @xmath4kpc ( depending on the adopted initial velocity distribution ) .
these catalogue can be used to test the hypothesis that some radio transients are related to these objects . |
the progress in single molecule tracking techniques provides a route to study the dynamical aspects of individual particles , such as colloids , polymers , vesicles or even cells .
specifically , insight is achieved in the influence of the environment on the particle behavior .
this is of particular importance for dilute suspensions , where hydrodynamic effects determine the transport properties of the solute particles . here , the solvent characteristics , e.g. , compressibility , are of paramount importance .
investigations of the dynamics of small objects suspended in a medium and driven by thermal fluctuations , i.e. undergoing brownian motion , have a long fruitful history @xcite . in the classical picture of brownian motion ,
the dynamics of a solute particle experiencing viscous drag and random collisions with the molecules of the solvent is described by the langevin equation , @xmath0 here @xmath1 is the particle mass , @xmath2 its coordinate , @xmath3 is the solvent s viscous friction coefficient and @xmath4 is a stochastic force which models collisions of the solute particle with the molecules of the solvent .
the stochastic force is _ assumed _ to be stationary , markovian , and gaussian with zero mean ( white noise ) , i.e. @xmath5 [ cols="^,^,^,^,^,^,^ " , ] at very short times , however , there is an apparent discrepancy between the simulations and the theory , as indicated in the insets of [ fig : rho10 ] and [ fig : rho5 ] . whereas simulations exhibit an exponential decay of the velocity autocorrelation function
, the theory predicts a decay which is faster than exponential .
this difference has a straightforward explanation . at very short times , the solvent in the simulations can not be treated as a continuous medium , and the navier - stokes equation can not be applied .
instead , one has to assume random collisions between solvent particles and the colloid . here , the langevin equation applies , and the velocity autocorrelation function indeed decays exponentially , with the decay constant given by the ratio of the enskog friction @xmath6 and the colloid mass @xmath1 @xcite , with @xmath7 this behavior is , in principle , a consequence of the coarse - graining approach of mpc - the relatively small number of solvent particles and the discrete dynamics imply that few collisions occur in a single simulation step , and , consequently , a continuum description is not applicable .
we remark that the enskog friction coefficient has a different physical origin than the stokes viscous friction coefficient , and that the two are not necessarily related .
the theory , on the other hand , assumes that a hydrodynamic description is valid down to arbitrarily short times .
the enskog regime is completely absent from the theoretical treatment .
consequently , the theoretical curves differ only when compressibility is accounted for , and only at the sonic time scale ( see [ fig : rho10 ] , right ) .
we further observe that for characteristic sonic times @xmath8 smaller than viscous times @xmath9 , the velocity autocorrelation functions decay monotonically and do not exhibit any peculiar properties .
however , when @xmath8 becomes larger than @xmath9 , anticorrelations clearly develop in the velocity autocorrelation function , signaling backtracking .
the monotonic decay of the vacf is superseded by crossing zero and an asymptotic approach of the hydrodynamic @xmath10 long - time tail .
here , the simulation results are in good qualitative agreement with the predictions of the theory .
we note that if , in the theoretical calculations , the effects of solvent compressibility are not taken into account , i.e. if the calculations are performed with an infinitely high velocity of sound , no anticorrelations develop ( dotted lines of [ fig : rho10 ] and [ fig : rho5 ] ) .
therefore , these anticorrelations are indeed a direct consequence of the interference of sonic and viscous effects .
velocity autocorrelation functions for solvent particles for @xmath11 and @xmath12 ( left ) and @xmath13 and @xmath14 ( right ) .
the points correspond to positive values of the vacf and the solid lines correspond to negative values of the vacf.[fig : solvent_vacf ] ] to study the influence of backtracking on long - time transport properties , we calculate the diffusion coefficient as the integral of the velocity autocorrelation function , i.e. , @xmath15 assuming the validity of the stokes - einstein relation , @xmath16 where @xmath17 , the product @xmath18 depends only on the properties of the colloid , namely its size , and does not depend on properties of the solvent , namely its shear viscosity , and serves as a good indicator of the relevance of sonic effects at long times . as shown in [ fig : rho10_diffusion ] and [ fig : rho5_diffusion ] , we observe that in the presence and in the absence of backtracking , @xmath19 tends to the same value in the asymptotic limit .
clearly , at the sonic time scale , there are significant deviations between the cases with and without backtracking . however , the long - time limit is always the same .
as before , the simulation results are in good qualitative agreement with theory .
we note that if the solvent compressibility is eliminated from the theoretical calculations , the long - time limit of @xmath19 remains the same ( dotted lines of [ fig : rho10_diffusion ] and [ fig : rho5_diffusion ] ) .
hence , we have confirmed that even when the effects of sound are very pronounced and dramatically influence the short - time regime of the solvent particles motion , the net effect remains zero , i.e. the role of sound can be neglected when studying long - time observables .
the main requirement to observe backtracking is that the sonic time scale should be equal to or slightly larger than the viscous time scale .
this is clearly not the case in typical conditions of a micron - sized particle suspended in a liquid like water since , as we demonstrated above , in this case the sonic time scale is of the order of nanoseconds , whereas the viscous time scale is of the order of microseconds .
however , the sonic time scale is proportional to the size of the particle , whereas the viscous time scale is proportional to the square of the size of the particle , and already for a particle a few nanometres in size , the sonic time scale and the viscous time scale have the same order of magnitude - namely , several picoseconds .
clearly , the viscous time scale can further be reduced by employing a more viscous fluid , and conditions under which backtracking occurs can be reached experimentally . to complete the analysis ,
we study the velocity autocorrelation function of the solvent itself @xcite . for small time steps , it is known to exhibit anticorrelations , and it may be tempting to assume that anticorrelations in the velocity of the solute are a direct consequence of anticorrelations in the velocities of the solvent particles
. however , we find that anticorrelations within the solvent and anticorrelations of the colloidal particle do not necessarily arise together , as can be seen by comparing the data for @xmath20 in [ fig : rho10 ] and [ fig : solvent_vacf ] ( left ) .
we have confirmed that interference of sound and viscous effects leads to a nonmonotonic decay of the velocity autocorrelation function of a single colloidal particle suspended in a viscous solvent .
furthermore , we have verified that under certain conditions backtracking can occur , i.e. the velocity autocorrelation function becomes negative .
this behavior is reminiscent of viscoelastic behavior - however , it has a very different , dynamical origin which lies in the interference of the formation of sound waves and hydrodynamic vorticity .
finally , we have confirmed that despite the strong influence of backtracking on the velocity autocorrelation function , its long - time integral , the diffusion coefficient , is unaffected .
that is to say that at long times the effects of sound are completely integrated out .
an important consequence of this result is the confirmation that sonic effects at the single - particle level can be neglected at long observation times .
mb and gf acknowledge financial support by the swiss national science foundation ( grant no . pp0022_119006 ) .
gf acknowledges prof .
gene stanley for several scientific and non - scientific discussions .
gompper , g. ; ihle , t. ; kroll , d. m. ; winkler , r. g. multi - particle collision dynamics - a particle - based mesoscale simulation approach to the hydrodynamics of complex fluids . _ advanced computer simulation approaches for soft matter sciences iii _
, _ advances in polymer science _ * 2009 * , 221 , 1 - 87 .
huang , c .- c . ; chatterji , a. ; sutmann , g. ; gompper , g. ; winkler , r. g. cell - level canonical sampling by velocity scaling for multiparticle collision dynamics simulations .
_ j. comput .
phys . _ * 2010 * , 229 , 168 . | the role of sound in the dynamics of mesoscale systems is typically neglected , since frequently the associated time scales are much smaller than all the other time scales of interest . however , for sufficiently small objects embedded in a solvent with a sufficiently small sound velocity , sound can play a crucial role .
in particular , behavior resembling viscoelasticity has been theoretically predicted for non - viscoelastic fluids .
this effect is due to the interference of the propagation of sound waves caused by the solute particle s motion and hydrodynamic vortex formation .
we demonstrate this effect , known as backtracking , in computer simulations employing the method of multiparticle collision dynamics .
we systematically study the influence of sound on the dynamics of the solute particle , and find that it disappears in the long - time limit .
thus , we confirm that sonic effects at the single - particle level can be neglected at sufficiently long times . |
the physical characteristics ( shapes , excitation energies , angular momentum , radial flow , ... ) of the hot nuclei formed in nucleus - nucleus collisions are obtained by comparisons between models and data .
statistical decay models are widely used and dedicated to describe the decay of hot fully equilibrated systems , defined as systems having reached energy , shape and isospin equilibrium @xcite .
however other experimental works show strong effects of the entrance channel on the decay modes of the formed hot systems @xcite .
these effects which are not taken into account in the statistical decay models can provide complementary information on the nuclei such as the characteristics of the nucleon - nucleon interaction @xcite and/or the viscosity of the nuclear matter .
up to now , these entrance channel effects have been studied by tracking deviations from the standard fission process in the quasi - projectile break - up @xcite .
such deviations have been seen on the angular distributions , which show the focusing of the break - up axis along the quasi - projectile velocity direction .
associated with this focusing , highly asymmetrical break - ups have also been observed , whereas symmetrical break - ups were expected .
the corresponding relative velocities were higher than the value predicted by the viola systematics @xcite .
these aligned break - ups represent up to 75% of the binary break - up for the xe+sn system @xcite . in this paper
, we will extend these studies performed for binary break - up to higher fragment multiplicities to test to what extent these effects are present for other exit channels .
we will present systematic studies on experimental data only , over a wide range of system sizes , incident energies and projectile / target asymmetries .
there will be no attempt to compare these data to theoretical calculations .
we have studied the decay of different quasi - projectiles formed in several projectile - target collisions from 25 to 90 @xmath0 . in the first section
, we present the experimental set - up , the studied sytems and the selections which have been applied on the data . in the second section
the charge distributions , the associated velocity and angular distributions obtained for the different systems are shown .
the third section is devoted to the evolutions of some experimental observables as a function of the break - up asymmetry of the quasi - projectile . in the last section our conclusions
are given .
the experiments were performed at the ganil facility with the indra detector . the different systems which will be presented in this article are ni+ni at 52 and 90 @xmath0 , xe+sn from 39 to 50 @xmath0 , ta+au at 33 and 39 @xmath0 , ta+u at 33 and 39 @xmath0 and u+u at 24 @xmath0 .
target thicknesses were respectively 193 @xmath1 for the experiment with the au target , 179 @xmath2 for the ni + ni experiment and 330 @xmath3 for the xe + sn experiment .
the 100 @xmath4 uranium target was deposited between two 20 @xmath4 carbon foils .
we will show afterwards the method used to separate the events corresponding to the uranium target from those corresponding to the carbon backing .
typical beam intensities were 3 - 4 @xmath5 pps .
events were registered when at least four charged particle detectors fired ( eight for the collisions with the ta projectile ) .
the indra detector @xcite can be schematically described as a set of 17 detection rings centered on the beam axis . in each ring
the detection of charged products was provided with two or three detection layers .
the most forward ring , @xmath6 , is made of phoswich detectors ( plastic scintillators ne102 + ne115 ) . between @xmath7 and @xmath8 eight rings
are constituted by three detector layers : ionization chambers , silicon and ics(tl ) . beyond @xmath8 ,
the eight remaining rings are made of double layers : ionization chambers and ics(tl ) .
the total number of detection cells is 336 and the overall geometrical efficiency of indra detector corresponds to 90% of 4@xmath9 .
isotopic separation is achieved up to z=3 - 4 in the last layer ( ics(tl ) ) over the whole angular range ( @xmath10 ) .
the charge resolution in the forward region ( @xmath11 ) is one unit up to z@xmath1250 . in the backward region ( @xmath13 )
a charge resolution of one unit is obtained up to z@xmath1220 .
the energy resolution is about 5% for ics(tl ) and ionization chambers and better than 2% for silicon detectors . and the velocity component parallel to the beam @xmath14 in the center of mass frame for the ta+au collisions at 39.6 @xmath0 and for events having a total linear momentum greater than 80shading of gray is darker for the high cross sections . ]
the binary character of the collisions is clearly seen on figure [ f : zv ] which plots the correlation between the charge @xmath15 and the velocity component parallel to the beam @xmath14 of each detected fragment .
the fragments are roughly distributed around two areas : one around the quasi - projectile velocity and charge ( @xmath16 and @xmath17 ) and one around the quasi - target velocity ( @xmath18 ) . due to the experimental thresholds ,
the fragments emitted by the quasi - target are not detected efficiently .
then , for this analysis , only the fragments produced by the quasi - projectile are taken into account . due to the quasi symmetry of most of the studied systems , a fragment ( z@xmath193 )
is assumed to be emitted by the quasi - projectile if its velocity component parallel to the beam axis is higher than the velocity of the center of mass @xmath20 .
in order to have enough information to perform our studies , only events for which more than 80% of the incident linear momentum was detected are selected ( @xmath21 where @xmath22 and @xmath23 are the charge of the fragment @xmath24 and the charge of the projectile respectively and @xmath25 and @xmath26 their velocity components parallel to the beam axis in the laboratory frame ) .
the events are sorted according to the multiplicity @xmath27 of fragments ( z@xmath193 ) emitted by the quasi - projectile .
the use of a @xmath28 multi - detector gives a very good coverage of the quasi - projectile emissions .
the angular , charge and velocity or energy distributions are precisely obtained .
this allows the study of the internal correlations in one event like the correlation between the fragments relative velocities and the break - up direction .
it will be shown later that this is a powerful tool to distinguish different reaction mechanisms . as said previously
, the uranium was deposited between two carbon layers .
thus , for the systems with this target , there will be a mixing of reactions on the backing ( c ) and on the target ( u ) .
the separation between these two reactions is done by using the angular distribution of the quasi - projectile . for a fixed incident energy , the grazing angle for the heavy target
is larger that the grazing angle for the light one .
the values of this angle are summarized in table [ t : grazing ] .
.grazing angles in the laboratory frame for different systems . [ cols="^,^,^ " , ] two contributions are then expected for the quasi - projectile diffusion angle @xmath29 .
the quasi - projectile ( qp ) is reconstructed with the fragments whose velocities are greater than @xmath20 .
such distributions are shown in figures [ f : tqp_tau33_all ] to [ f : tqp_uu24_all ] .
in the first two figures , the dashed line corresponds to the system with the ta projectile and the au target and the full line corresponds to the system with the ta projectile and the u target and c backing .
reactions on the c backing clearly show up at small angles , while the distributions at large angles , from their similarities with those arising from ta+au , can be attributed to reaction on u only . a clear selection of ta+u reactions is ensured by selecting events with @xmath30 . for the u+c , u system ( figure [ f : tqp_uu24_all ] ) ,
the two contributions ( u target and c backing ) are better separated due to the larger difference between the values of the grazing angles .
the same cut ( @xmath31 ) was applied to separate the contributions from the uranium target and the carbon backing . for this system ,
the contribution of the uranium target at small angles is weaker compared to the ta+c , u systems . for the ta+c , u system at 33 @xmath0 ( full line ) and the ta+au system at 33 @xmath0 ( dashed line ) .
] for the ta+c , u system at 39.6 @xmath0 ( full line ) and the ta+au system at 39.6 @xmath0 ( dashed line ) . ] for the u+c , u system at 24 @xmath0 . ] and @xmath29 . ]
( lowermost row ) distributions for the ta+au system at 33 @xmath0 .
the columns correspond to different fragment mutiplicities ( from @xmath27=2 to @xmath27=4 from left to right ) .
the shading of the distribution is darker and darker according to the rank of the fragment in the event ( the lightest shading correspond to the heaviest fragment in the event ) . ] for the ta+au system at 39.6 @xmath0 . ] for the xe+sn system at 39 @xmath0 . ] for the xe+sn system at 45 @xmath0 . ] for the xe+sn system at 50 @xmath0 . ] for the ni+ni system at 52 @xmath0 . ] for the ni+ni system at 90 @xmath0 . ] for the ta+u system at 33 @xmath0 ( @xmath32 ) . ] for the ta+u system at 39.6 @xmath0 ( @xmath32 ) . ] for the ta+c , u system at 33 @xmath0 ( all @xmath29 ) . ] for the ta+c , u system at 39.6 @xmath0 ( all @xmath29 ) . ] for the u+u system at 24 @xmath0 ( @xmath32 ) . ] for the u+c system at 24 @xmath0 ( @xmath33 ) . ]
we will first look at simple and direct observables like the fragment charge distributions , the corresponding distributions of the velocity component parallel to the beam ( @xmath14 ) and the angular distributions for the different systems .
the angle @xmath34 is the angle between the velocity of the fragment in the quasi - projectile ( qp ) frame and the velocity of the quasi - projectile in the center of mass frame ( see figure [ f : angles ] ) . the events are sorted according to the multiplicity , @xmath27 , of the forward fragments ( @xmath35 ) . in each event , the fragments are sorted according to their charge .
charge , @xmath14 and @xmath36 distributions are shown on figure [ f : hier_taau33 ] .
the uppermost row corresponds to the charge distribution , the middle row to the @xmath14 distribution and the lowermost row to @xmath36distribution .
the different columns correspond to the different fragment mutiplicities ( from @xmath27=2 to @xmath27=4 from left to right ) . on each panel
, the shading of the distribution darkens according to the rank of the fragment in the event .
one spectacular observation is that the velocity distributions are strongly correlated to the charge sorting : the heaviest fragment is in average the fastest one , the second heaviest fragment is the second fastest one and so on . for the @xmath36 distributions ,
the distribution of the heaviest fragment is peaked at forward angles ( @xmath36@xmath37 ) : the emission direction of this fragment is close to the quasi - projectile velocity direction .
such pictures were not expected in case of a fragmentation of a fully equilibrated quasi - projectile , for which the fragments are emitted in all directions without a particular hierarchy for the velocities .
this hierarchy effect ( the ranking in charge induces on average the ranking in @xmath14 and the @xmath36 distribution of the heaviest fragment is forward - peaked ) suggests an entrance channel effect .
similar results are observed for the corresponding distributions for most of other systems .
the heaviest fragment is the fastest one and its direction of emission is close to the quasi - projectile velocity direction whatever the system size and whatever the incident energy .
the hierarchy between the size of the fragment and its velocity is observed for the ta + au ( figures [ f : hier_taau33 ] and [ f : hier_taau39 ] ) , xe+sn ( figures [ f : hier_xesn39 ] , [ f : hier_xesn45 ] and [ f : hier_xesn50 ] ) and ni+ni systems ( figures [ f : hier_nini52 ] and [ f : hier_nini90 ] ) .
this effect is strong on the ta+au system and is slightly weaker for the xe+sn and the ni+ni systems . for the xe+sn system ,
the anisotropy of the heaviest fragment angular distribution increases slightly with the incident energy . for the ta+u system ( events with @xmath30 , figures [
f : hier_tau33 ] and [ f : hier_tau39 ] ) , the same hierarchy effect is observed as for the ta+au system . for the ta + c , u system ( figures [ f : hier_tau33_all ] and [ f : hier_tau39_all ] ) , this hierarchy effect is still observed for high multiplicities ( @xmath27@xmath38 ) , but is not present for @xmath27=2 .
this is due to the mixing between the carbon backing and the uranium target mentioned in the previous section . for @xmath27=2 ,
the two fragments have almost the same charges , the velocities are similar and the whole range of @xmath36is covered .
the contribution of the ta+c collisions is stronger and stronger when @xmath27 decreases .
this contribution is especially seen for @xmath27=2 ( compare leftmost columns of figures [ f : hier_tau33 ] and [ f : hier_tau33_all ] for 33 @xmath0 and figures [ f : hier_tau39 ] and [ f : hier_tau39_all ] for 39 @xmath0 ) where the distributions are completely different .
the ta+c collisions lead to the formation of an incomplete fusion nucleus , which decays through fission . in this case
, no hierarchy in velocities or no privileged angles of the heaviest fragment are expected . for the u+u system ( figure [ f : hier_uu24 ] , events with @xmath39 ) the hierarchy effect is observed for @xmath27@xmath40 but is not observed for @xmath27=2 . due to the high fissility of uranium , the @xmath27=2 events correspond mainly to the fission process .
however , events for which the charge of the heaviest fragment is close to 80 are still present .
they correspond to asymmetric binary break - ups which are not expected for the fission of such a heavy nucleus .
these asymmetric break - ups are not present for the u+c system ( figure [ f : hier_uc24 ] , leftmost column , @xmath41 ) and no hierarchy effect is seen : these observations are consistent with the fission of a fully equilibrated nucleus . for @xmath27@xmath40 ,
the velocities of the two heaviest fragments are similar and their @xmath36 distributions are flat . for the other fragments ,
the hierarchy effect is present .
this could be due to a two - step process in which the hierarchy effect is present at the early stage of the collision , followed by the fission of the heaviest fragment due to the high value of its charge . to summarize this section , a hierarchy effect ( the ranking in charge induces the ranking in @xmath14 and
the @xmath36 distribution of the heaviest fragment is forward - peaked ) is seen for many system sizes and incident energies .
these observations are not consistent with the decay of a fully equilibrated nucleus for which no hierarchy effect is expected .
this `` hierarchy effect '' is consistent with a strong deformation of qp or qt during the collision which is followed by the break - up of these elongated nuclei in two or more fragments ( neck formation and break - up , see figure [ f : schcol ] ) .
the fragments emitted by this neck reflect its internal structure : its size at the center of mass is on average thinner than close to the quasi - projectile or quasi - target , and the velocity modulus in the center of mass frame of the nucleons in the neck close to the center of mass are smaller than those of nucleons close to the quasi - target or to the quasi - projectile .
this leads to the `` hierarchy effect '' observed experimentally .
the velocity gradient can be tracked by studying the relative velocities between the heaviest fragment and the others . in the proposed scenario , this relative velocity is expected to be larger than the one obtained from the decay of a fully equilibrated nucleus . for the heaviest systems , the standard fission process superimposes on the neck formation and break - up .
the two contributions are clearly seen for ta and u projectiles . in these cases , deviations from the standard fission
can be easily tracked .
in the previous section , we have evidenced a `` hierarchy effect '' , indicating a strong entrance channel influence on the fragmentation of the quasi - projectile .
we will study in this section correlations between some observables to verify whether the fragmentation scenario proposed above resists a more detailed analysis .
if an elongated quasi - projectile is formed and breaks quickly , the relative velocities between the fragments should be higher along the qp velocity direction than in any other direction , because the incident energy is not fully damped in internal degrees of freedom in such a process . within this assumption ,
the heaviest fragment focused along the qp velocity has a high velocity because it is a remnant of the projectile . in the case of the fragmentation of a fully equilibrated qp
, the relative velocity should be independent of the break - up direction . for @xmath27=2 ,
the relative velocity between the two fragments can be easily determined , while it is not the case for higher multiplicities .
we have seen in the previous section that the heaviest fragment has the same properties whatever the value of @xmath27 : it is the fastest one and it is focused in the direction of the quasi - projectile in the center of mass frame .
we can use this property to define the break - up direction @xmath42 as the direction of the heaviest fragment with respect to the quasi - projectile velocity , and the relative velocity @xmath43 as the difference of the velocity of the heaviest fragment @xmath44 and the velocity of the center of mass of all other fragments detected with a velocity greater than @xmath20 .
they are determined as follows : @xmath45 where @xmath22 is the charge of the @xmath46 fragment ( @xmath47 ) and @xmath48 its velocity @xmath49 is the velocity of the quasi - projectile .
all the velocities in formulae 1 - 3 are expressed in the center of mass frame . with this definition ,
@xmath42 is identical to @xmath34 of the heaviest fragment ( see figure [ f : angles ] ) .
let us study the behaviour of the @xmath51 distributions .
the experimental distributions are shown on figures [ f : asctp_taau33 ] and [ f : asctp_taau39 ] for the ta + au system at 33 and 39.6 @xmath0 respectively . on each figure
, the columns correspond to different values of @xmath27 and the rows to different ranges in charge of the heaviest detected fragment @xmath52 .
this cut in @xmath52 was done due to the particular behavior of the heaviest fragment .
it is also correlated to the asymmetry of the break - up as in fission : the highest values of @xmath52 correspond to the more asymmetric break - ups ( one big fragment and small other ones ) and the lowest values of @xmath52 correspond to the most symmetric break - ups ( equal size fragments ) .
it can be seen on these figures that the @xmath51 distributions are in most cases peaked at 1 ( emission along the direction of the qp velocity ) .
the most forward peaked @xmath51 distributions are obtained for the most asymmetric break - ups ( high values of @xmath52 ) and this for all values of @xmath27 .
this is not expected in the case of a fragmentation of a fully equilibrated nucleus .
an angular momentum effect would lead to a forward - backward symmetry in angular distributions .
for the most symmetric break - ups ( low values of @xmath52 ) and the greater imf multiplicities , the distribution becomes less forward peaked . in this case , the fragments having almost the same charge , the ranking in charge is meaningless .
consequently the memory of the direction of the qp in the fragmentation process is blurred .
another possible scenario could be the occurrence of the fragmentation of a fully equilibrated qp .
it will be seen later that the cut at @xmath20can also influence these distributions .
distributions for the ta+au system at 33 @xmath0 .
the columns correspond to the different fragment multiplicities from 2 to 5 .
the rows correspond to different ranges for the charge @xmath52 of the heaviest fragment : the most symmetrical break - ups ( low @xmath52 values ) correspond to the uppermost row and the most asymmetrical break - ups ( high @xmath52 values ) correspond to the lowermost row . ] for the ta+au systeme at 39.6 @xmath0 . ]
is there a connexion between the strength of the anisotropy of the @xmath51 distributions and the violence of the collisions ?
the latter has been estimated by using the sum of the transverse energies @xmath53 of light charge particles ( @xmath54 ) .
this global variable has been widely used as an impact parameter selector in previous studies @xcite .
the low values of transverse energies correspond to peripheral collisions and the high values to the most violent collisions .
the reduced transverse energy distributions ( @xmath55 where @xmath56 is the energy available in the center of mass ) are presented in figure [ f : aset12_taau39 ] for ta+au collisions at 39.6 @xmath0 .
the shapes of these distributions depend weakly on the value of @xmath27 and strongly on that of @xmath52 .
this is in agreement with @xmath57 being the remnant of the qp .
the average @xmath53 value decreases when going from symmetrical ( first row ) to asymmetrical break - ups ( last row ) as already shown for the lighter ar+ni system @xcite .
the last row corresponds to the most asymmetrical and aligned break - ups ( see figure [ f : asctp_taau39 ] ) .
the transverse energies of the associated particles are low and correspond to peripheral collisions .
the first row corresponds to the most symmetrical break - ups and to more wider @xmath51 distributions .
the transverse energies of the associated particles are high and correspond to more central collisions .
it is interesting to observe low @xmath53 values for the highest imf multiplicities ( last row , rightmost panel of figure [ f : aset12_taau39 ] ) . in this case , the excitation energy of the qp is low and low evaporated imf multiplicities are expected .
this observation , combined with the `` hierarchy effect '' and the strongly peaked @xmath51 distributions , strongly suggests that these fragmentations are only due to entrance channel effects .
for the binary break - ups ( first column ) , the very asymmetrical break - ups ( last row ) are not expected for the standard fission of a quasi - ta . nevertheless , this is the dominant decay channel for @xmath27=2 .
the second and third panels of the first column ( @xmath27=2 ) correspond to more symmetrical break - ups ( @xmath58 ) and the standard fission process is expected . in this case
the transverse energy distributions are wider .
it seems that two contributions could be considered as already suggested by the the @xmath51 distributions .
these observations are also made on the other systems .
distributions ( in percent ) for the ta+au system at 39.6 @xmath0 .
the columns correspond to the different fragment multiplicities from 2 to 5 .
the rows correspond to different ranges for the charge @xmath52 of the heaviest fragment : the most symmetrical break - ups ( low @xmath52 values ) correspond to the uppermost row and the most asymmetrical break - ups ( high @xmath52 values ) correspond to the lowermost row . ] with @xmath51 for the ta+au system at 39.6 @xmath0 .
the columns correspond to the different fragment multiplicities from 2 to 5 .
the rows correspond to different ranges for the charge @xmath52 of the heaviest fragment : the most symmetrical break - ups ( low @xmath52 values ) correspond to the uppermost row and the most asymmetrical break - ups ( high @xmath52 values ) correspond to the lowermost row . ] for the u+u system at 24 @xmath0 ( @xmath32 ) . ] for the u+c system at 24 @xmath0 ( @xmath33 ) . ] as a fonction of @xmath51 for the ta+au system at 39.6 @xmath0 and for @xmath27=2 .
the two other columns display the correlation between @xmath59 and @xmath51 for two different systems .
see text for more details . ]
let us now come to the correlation between the relative velocities of the fragment @xmath59 and @xmath51 .
they are presented on figure [ f : vrctp_taau39 ] for ta+au collisions at 39.6 @xmath0 , on figure [ f : vrctp_uu24 ] for the u+u collisions at 24 @xmath0 and on figure [ f : vrctp_uc24 ] for the u+c collisions at 24 @xmath0 .
as in figures [ f : asctp_taau33 ] and [ f : asctp_taau39 ] , the columns correspond to different imf multiplicities and the rows to different values of @xmath52 . for the most asymmetric breakup of the ta + au collisions ( two lowest rows for figure [ f : vrctp_taau39 ] ) , a strong variation of @xmath59 with @xmath51
the value of @xmath59 increases from 2.3 @xmath60 at @xmath51=-1 up to @xmath61 at around @xmath62 and then decreases . the variation is weaker for the more symmetric break - ups ( two first rows ) .
this variation is also observed for the u+u collisions but its amplitude is smaller . for the u+c system ( figure [ f : vrctp_uc24 ] ) , no evolution is seen whatever the imf multiplicity and the charge of the heaviest fragment . for the u+c system ,
no `` hierarchy effect '' was seen ( see figure [ f : hier_uc24 ] ) . within our interpretation of the `` hierarchy effect ''
, no modulation is expected for this system . on all other systems ,
the same correlation is observed : a modulation of @xmath59 with @xmath51 is seen when the `` hierarchy effect '' is observed .
so the modulation of @xmath59 with @xmath51 is correlated to the observation of a `` hierarchy effect '' .
the decrease of @xmath59 at the smallest angles ( @xmath51=1 ) is not compatible with the expected behaviour of @xmath59 resulting from our interpretation on the `` hierarchy effect '' .
but due to the selection of the fragments in this analysis ( @xmath63 ) , the highest values of @xmath59 can not be reached .
we have tested the effect of this velocity cut on our analysis with the help of a simple simulation .
for sake of simplicity , we have only verified the effect of the velocity cut for @xmath27=2 .
the basic ingredients of this simulation are the following : a qp with a charge @xmath64 , a velocity @xmath65 and an angle @xmath29 is considered .
this qp splits in two fragments of charge @xmath52 and @xmath66 , the axis of the break - up being isotropically distributed and the relative velocity between the fragments ranging from 0 cm / ns to 10 cm / ns . for each case , the event is taken into account in this analysis if both fragments have a @xmath14 greater than @xmath20 .
the velocities @xmath44 and @xmath67 of the two fragments are calculated the following way : @xmath68 @xmath69 the first column of figure [ f : seuils_vrctp ] corresponds to the result of the simulation ; the second column shows the experimental bidimensional plots of @xmath59 versus @xmath51 for the ta + au collisions at 39.6 @xmath0 ; the third column shows the same plots for the u + c collisions at 24 @xmath0 ( @xmath41 ) .
the plots corresponding to events with an imf multiplicity equal to 2 are displayed .
the rows correspond to different selections on the charge of the heaviest fragment @xmath52 , the uppermost row corresponding to the lowest @xmath52 values and the lowermost row to the highest @xmath52 values . for each plot of the simulation ( leftmost column ) , the values of @xmath65 , @xmath29 and the ratio @xmath70 have been chosen according to the average experimental values for the ta+au collisions at 39.6 @xmath0 for a given imf multiplicity and for a fixed @xmath52 range .
the shaded areas ( leftmost column ) correspond to the events selected in our analysis .
one can see that the velocity cut ( @xmath63 ) systematically removes the events with the highest @xmath59 values when the fragmentation axis is aligned on the qp velocity direction . for @xmath51=1 ,
this effect is always present .
this could explain the decrease of @xmath59 for high @xmath51 values observed in figures [ f : vrctp_taau39 ] and [ f : vrctp_uu24 ] .
this is confirmed by the bidimensional plot ( second column ) which was used to obtain figure [ f : vrctp_taau39 ] .
the effect of the velocity cut is clearly seen for the highest @xmath52 values .
the agreement between data and simulation is good , showing that the velocity cut limits the range of accessible @xmath59 values . for higher @xmath27 values ,
similar conclusions are obtained .
another interesting observation is that the @xmath59 distribution for @xmath51@xmath71 is narrow , its average value @xmath59@xmath72 corresponds to the viola systematics @xcite and is not affected by the velocity cut at @xmath20 . when @xmath51 increases , the width of the @xmath59 distribution and its average value increase while the velocity cut has no influence , i.e. for @xmath51 values below 0.8 . above @xmath51=0.8 , high values of @xmath59
can not be reached and then its average value decreases .
for the u+c system rightmost column , it can be seen that the @xmath59 distributions are always narrow and centered around the viola systematics value whatever the @xmath51 value . for these collisions
, our experimental cut does not affect the evolutions of @xmath59 with @xmath51 .
the observations made in this section confirm the scenario proposed in the previous section . for the u+c system ,
all observations are compatible with the decay of a fully equilibrated nucleus : no modulation of @xmath59 with @xmath51 is seen while at the same time the `` hierarchy effect '' is absent or very weak . for the systems with a heavy target , the variation of @xmath59 with @xmath51
is observed .
this indicates that a strongly deformed qp or a neck of matter is formed at the early stage of the collision , and breaks quickly into several fragments .
in such a process , the incident kinetic energy is not fully damped , and the fragments keep the memory of the formation of the neck and/or deformed qp . within this interpretation ,
the heaviest fragment is a remnant of the projectile .
this scenario seems to occur for a wide range of system size and incident energies .
for the heaviest systems , like u+u for which the fissility of the qp is high , it competes with the fission process .
the study of the parallel velocity distributions of the fragments forward emitted as a function of their charge exhibits a `` hierarchy effect '' : the ranking in charge induces in average a ranking in the average parallel velocity , the heaviest fragment is the fastest and is focused in the forward direction relative to the qp recoil velocity .
this `` hierarchy effect '' is stronger when the size of the target is large , when the fissility of the qp is limited ( @xmath73 ) and when the incident energy is high .
this effect is observed for all systems , whatever the multiplicity of fragments .
these observations are compatible with the formation of a neck of matter in - between the quasi - projectile and the quasi - target
. the neck may be or not attached to the qp ( or qt ) remnant . in any case , its break - up is fast enough to keep the memory of the entrance channel .
this mechanism is predominant for asymmetric break - ups which mainly correspond to peripheral collisions . for the most peripheral collisions , the asymmetry observed for the binary break - ups
is not compatible with the result of a standard fission , and for the highest multiplicities , the high focusing of the heaviest fragment at forward angles is inconsistent with the fragmentation of a fully equilibrated nucleus .
the observed variation of the fragments relative velocity with the emission angle of the heaviest fragment strenghtens the previous statement .
this internal correlation study is a powerful tool to clearly establish the degree of equilibration of the studied ensemble . up to now , the comparisons of models to data were mainly made on global observations only ( size distributions , multiplicity distributions , kinetic energy distributions , and so on ... ) .
the study of such internal correlations is a more accurate tool to test the pertinence of the models . in order to have a full understanding of the fragmentation process
, the models have to reproduce the effects of the entrance channel .
a first step could be to mock up these effects in effective parameters in the statistical decay approaches
. it will be certainly worth to use dynamical approaches @xcite to test the presence of such a hierarchy effect .
but in all cases , the models have to reproduce the global observations and the internal correlations as well . | the fragmentation of the quasi - projectile is studied with the indra multidetector for different colliding systems and incident energies in the fermi energy range .
different experimental observations show that a large part of the fragmentation is not compatible with the statistical fragmentation of a fully equilibrated nucleus .
the study of internal correlations is a powerful tool , especially to evidence entrance channel effects .
these effects have to be included in the theoretical descriptions of nuclear multifragmentation . |
acetylene constitutes a key ingredient in the production of large complex hydrocarbon molecules in the dense interstellar medium ( herbst 1995 ) .
because acetylene has no permanent dipole moment , it lacks a rotational spectrum that could be observed at radio wavelengths ; observations of interstellar acetylene have therefore been limited to mid - infrared studies of rovibrational bands , carried out from ground-(e.g . , evans , lacy & carr 1991 ; carr et al . 1995 ) and space - based ( e.g. , lahuis & van dishoeck 2000 ; boonman et al .
2003 ; lahuis et al .
2007 ) observatories .
acetylene has been detected in the gas - phase - either in absorption ( e.g. , carr et al .
1995 ; lahuis & van dishoeck 2000 ) or in emission ( e.g. boonman et al .
2003 , this paper ) - mostly toward young stellar objects .
c@xmath0h@xmath0 can be used as a tracer of warm ( 100 k to 1000 k ) molecular gas along often complicated sightlines .
c@xmath0h@xmath0 abundance estimates , which were sometimes a few orders of magnitude higher than the predictions of cold gas - phase steady - state chemical models , have led to a better understanding of the role that warm gas - phase chemistry ( e.g. , doty et al .
2002 ; rodgers & charnley 2001 ) and/or grain mantle processing ( e.g. , ruffle & herbst 2000 ) can play in star - forming regions , both locally and in extra - galactic objects ( lahuis et al . 2007 ) . in this letter , we present the first detection of the @xmath1 band of acetylene ( c@xmath0h@xmath0 ) at 13.7 @xmath2 m toward the star forming region cepheus a east using the infrared spectrograph ( irs ) onboard the _ spitzer space telescope_. this is the first map of c@xmath0h@xmath0 obtained toward any interstellar gas cloud .
section 2 describes the observations and data analysis .
sections 3 and 4 compare the spatial distribution of c@xmath0h@xmath0 to those of gaseous co@xmath0 and h@xmath0 @xmath11(2 ) , and discuss the c@xmath0h@xmath0-emitting gas in the context of shock chemistry and local outflow activity .
the presence of c@xmath0h@xmath0 on interstellar dust grains will also be discussed in the context of cometary ices composition .
spectral maps were obtained with the irs instrument onboard _ spitzer _ as part of the guaranteed time observer program 113 .
1@xmath12@xmath31@xmath12 square fields were observed with the short - low ( sl ) , short - high ( sh ) , and long - high ( lh ) modules , providing wavelength coverage from 5.2 to 25 @xmath2 m .
we obtained continuous spatial sampling by stepping the slit perpendicular and parallel to its long axis in steps of one - half its width and 4/5 its length , respectively .
the data were processed with version 12 of the pipeline .
we used the spectroscopy modeling analysis and reduction tool ( smart , higdon et al .
2004 ) , along with locally developed routines ( neufeld et al .
2006 ) , to extract wavelength- and flux - calibrated spectra and to generate spectral line intensity maps . to estimate the uncertainties in the c@xmath0h@xmath0 line intensities ( @xmath1 band at 13.7 @xmath2 m )
, we first calculated the standard deviation ( @xmath13 ) around our best fit to the local continuum for each point in the map .
we then shifted the best - fit continuum by @xmath141 @xmath13 and generated the corresponding new estimates of the line intensities .
the difference between the best - fit line intensities and the intensities obtained by shifting the best - fit continuum model constitutes our @xmath141 @xmath13 errors .
while extinction due to the long wavelength wing of the silicate feature ( 9.7 @xmath2 m ) certainly occurs , corrections to the c@xmath0h@xmath0 line intensities were not applied . as we will see below
, c@xmath0h@xmath0 arises in a warm component located mostly in front of the cold quiescent gas traced by the silicates ( sonnentrucker et al .
2007 , apj in press ) making a direct estimate of the fraction of dust located in this warm component nearly impossible . since our study also indicates that the silicates exhibit a homogeneous distribution over the spatial extent of the gaseous c@xmath0h@xmath0 emission , we are confident that the c@xmath0h@xmath0 intensity variations we see are not predominantly due to extinction effects .
figure 1 ( _ upper _ ) shows a summed irs spectrum in the wavelength range containing acetylene at the spatial location corresponding to one peak in the c@xmath0h@xmath0 line emission intensity map ( hw5/6 in fig . 2 ) .
note that frequent order mismatches around 14 @xmath2 m preclude any reliable detection of the hcn @xmath15 band at 14.05 @xmath2 m with our data .
the striking similarities between the c@xmath0h@xmath0 maps presented here ( fig . 2 ) and the gaseous co@xmath0 maps we found previously ( sonnentrucker et al .
2006 ) strongly suggest that c@xmath0h@xmath0 arises in the same warm postshock gas component exhibiting gaseous co@xmath0 . because the c@xmath0h@xmath0 emission lines are weaker by a factor @xmath510 than those of co@xmath0 , we first searched for the spatial positions where c@xmath0h@xmath0 was detected at a @xmath161.5 @xmath13 level .
38 such individual spatial positions fulfilled this criterion over the mapped region shown in fig . 2 .
these are the spectra that we will consider when comparing the c@xmath0h@xmath0 distribution with that of gas - phase co@xmath0 and h@xmath0 @xmath11(2 ) , a tracer of shock activity ( see fig . 3 ) .
we constrained the gas temperature in the selected c@xmath0h@xmath0-emitting region , by comparing synthetic profiles of the @xmath1 acetylene band for temperatures ranging from 50 to 900 k with the observed c@xmath0h@xmath0 @xmath17-branch ( 13.71 @xmath2 m ) profile .
we find that temperatures between 50 and 200 k best fit the 38 selected c@xmath0h@xmath0 line profiles ( fig . 1 , _ lower
note that significantly higher temperatures , such as those found by lahuis & van dishoeck ( 2000 ) toward other massive young stellar objects ( 300 - 1000 k ) , are ruled out in cepheus a east .
most importantly , our moderate 50 - 200 k range coincides with the temperature range found for gaseous co@xmath0 in that same region ( sonnentrucker et al .
2006 ) arguing in favor of a common spatial origin for the two species . to estimate the column density associated with the c@xmath0h@xmath0 emission
, the dominant excitation mechanism needs to be identified .
accurate collisional rates for vibrational excitation of acetylene are not available to our knowledge . if the collisional rates for c@xmath0h@xmath0 vibrational excitation were similar to those of co@xmath0 ( boonman et al . 2003 ) , densities higher than @xmath18 @xmath19 would be required to account for the observed emission .
such a high density would give rise to strong high - lying h@xmath0o rotational lines that are , however , not detected ( e.g. , van den ancker et al . 2000 ) .
additionally , considering the low temperature in the c@xmath0h@xmath0-emitting gas component relative to the energy of the @xmath20 state ( @xmath21 k ) , as well as the rather low local hydrogen density inferred over this region ( @xmath22@xmath5 few @xmath3 10@xmath23 to a few @xmath3 10@xmath24 @xmath19 ; codella et al .
2005 ) , it is quite unlikely that collisions are the dominant excitation mechanism .
radiative pumping by 13.7 @xmath2 m continuum photons produced by warm dust local to c@xmath0h@xmath0 is also unlikely to produce the observed emission , since we previously determined that the radiation field in this region is dominated by dust close to the hw2 protostellar region ( sonnentrucker et al .
therefore , we conclude that radiative pumping by 13.7 continuum photons emanating from the hw2 protostellar region seems the most likely excitation mechanism for the observed acetylene molecules , as we also found for co@xmath0 ( sonnentrucker et al .
2006 ) .
we computed the c@xmath0h@xmath0 column densities over the mapped region , following the prescription described in sonnentrucker et al .
( 2006 ) . under these assumptions , the derived @xmath10(c@xmath0h@xmath0 )
values are proportional to the acetylene intensities and to the square of the angular separation from the exciting source , hw2 ( e.g. , gonzlez - alfonso et al . 2002 ) .
figure 2 displays the intensity map ( _ upper _ ) and the derived column density map ( _ lower _ ) for c@xmath0h@xmath0 toward cepheus a east . the gray contours show the distribution of nh@xmath25 ( 1,1 ) a tracer of cold quiescent molecular gas ( torrelles et al .
1993 ) toward the region fully sampled by _
spitzer_. like gas - phase co@xmath0 , the @xmath17-branch of gaseous c@xmath0h@xmath0 mainly traces the walls of the cavity carved by the _ northeast _ component of the outflow originating at hw2 ( e.g. , gmez et al . 1999 ) , an outflow that is apparently responsible for the disruption of the quiescent clouds as seen in the nh@xmath25 distribution ( e.g. , torrelles et al . 1993 ; goetz et al .
1998 ; van den ancker 2000 ) .
intensity maxima for c@xmath0h@xmath0 are detected close to the hw6 radio continuum source , at the ne bridge that joins the cep a-2 and cep a-3 cloud cores ( torrelles et al . 1993 ) , and along the western surface of the cep a-3 cloud .
weaker emission is also found further from the ne bridge .
the similarities in both the temperature and the spatial distribution of the c@xmath0h@xmath0 and co@xmath0 emitting gas add weight to the conjecture that both species arise from the same warm shocked component .
we finally searched for correlations between the acetylene measurements and those obtained for gas - phase co@xmath0 and h@xmath0 @xmath11(2 ) , a tracer of shock activity , for those 38 positions where c@xmath0h@xmath0 was detected at a @xmath161.5@xmath13 level .
figure 3 compares the c@xmath0h@xmath0 intensity with that of gaseous co@xmath0 ( _ upper _ ) , as well as the c@xmath0h@xmath0 column density with the co@xmath0 column density ( _ middle _ ) and the h@xmath0 @xmath11(2 ) intensity ( _ lower _ ) . in all cases
, the good correlation found between acetylene , gaseous co@xmath0 and h@xmath0 @xmath11(2 ) indicates that c@xmath0h@xmath0 does indeed arise in the warm shocked component also containing gaseous co@xmath0 .
acetylene was previously observed toward low - to - high mass star - forming regions either in absorption ( e.g. , lahuis & van dishoeck 2000 ) or in emission ( e.g. , boonman et al .
2003 ) using the _ infrared space observatory _ ( _ iso _ ) .
excitation temperatures ranging from @xmath5 10 to 900 k and abundances with respect to h@xmath0 ranging from a few @xmath3 10@xmath9 to a few @xmath3 10@xmath26 were derived .
steady - state models of gas - phase chemistry in cold ( 10 - 50k ) dense ( 10@xmath27 - 10@xmath28 @xmath19 ) molecular clouds predict abundances for acetylene between a few @xmath3 10@xmath29 and 1@xmath3 10@xmath9 with respect to h@xmath0 depending on the role that neutral - neutral destruction reactions may play ( bettens , lee & herbst 1995 ; lee et al .
similar abundances are predicted by models of gas - grain chemistry in quiescent clouds with the highest values obtained only after 10@xmath30 years ( ruffle & herbst 2000 ) .
while such models could account for the observed abundances in the cold gas , they were unable to reproduce the enhancements observed toward much warmer gas components in those objects .
mechanisms such as c@xmath0h@xmath0 ice sublimation from grain mantles and/or c@xmath0h@xmath0 enhancements via warm gas - phase chemistry were then invoked ( e.g. , carr et al . 1995 ; doty et al .
2002 ; boonman et al .
2003 ) . for the ne outflow region in cepheus a east
, we derive c@xmath0h@xmath0/h@xmath0 abundance ratios in the range 1 @xmath3 10@xmath8 to 4 @xmath3 10@xmath9 , for an assumed h@xmath0 column density of 1.5 @xmath3 10@xmath31 @xmath19 ( gmez et al . 1999 ) .
these values are averages along the observed sight - lines .
the highest abundances are localized around and to the south of , the ne position as well as around the positions of the hw5/6 sources ( see fig . 2 ) .
this is precisely where the interaction between the _ northeast _ outflow and the ambient molecular clouds occurs , and it is in these regions that we observe strong h@xmath0 @xmath11(2 ) emission , a tracer of warm shocked gas .
thus the spatial variation of the c@xmath0h@xmath0 abundance again strongly suggests an association with shock activity , perhaps as a result of ( 1 ) production in the gas phase via high temperature reactions , or ( 2 ) grain mantle sputtering . models for chemistry in hot cores ( rodgers & charnley 2001 ; doty et al . 2002 ) indicate that enhanced abundances of c@xmath0h@xmath0 ( @xmath5 few @xmath3 10@xmath9 ) are expected in warm regions with @xmath32 200 k. the good correlations between h@xmath0 @xmath11(2 ) , gaseous co@xmath0 and c@xmath0h@xmath0 indicate that such high temperatures were reached in the warm gas at the passage of the non - dissociative shock .
the chemical pathways leading to the production of c@xmath0h@xmath0 are slow , and enhanced abundances only occur after @xmath5 10@xmath33 years , a time scale much greater than that expected for shock heating of the gas ( @xmath5 300 years ; e.g. kaufman & neufeld 1996 ) .
hence , enhanced production of c@xmath0h@xmath0 by high temperature gas - phase chemistry is unlikely to be predominant in the observed region .
thus the correlations shown in fig . 3 argue in favor of grain mantle sputtering over gas - phase production as the origin of the c@xmath0h@xmath0 in cepheus a east .
this is the same production mechanism that we favored for gaseous co@xmath0 ( sonnentrucker et al .
while models for the production of c@xmath0h@xmath0 in shocks are not available to our knowledge , our results further suggest that both c@xmath0h@xmath0 and co@xmath0 are released into the gas phase under very similar physical conditions .
the gaseous @xmath34 ratio is roughly constant , with a mean value of 0.08 for all sight - lines where we detected acetylene at the 1.5 @xmath13 level . if this value reflects the composition of the grain mantle , and given a @xmath10(co@xmath0)@xmath35/@xmath10(h@xmath0o)@xmath35 ratio in this source of 0.22 ( sonnentrucker et al .
2007 , apj in press ) , then the required @xmath10(c@xmath0h@xmath0)@xmath35/@xmath10(h@xmath0o)@xmath35 ratio is 0.02 .
this value is at least a factor 4 larger than those derived toward other star - forming regions ( 0.1 - 0.5% , evans et al . 1991 ; lahuis & van dishoeck 2000 ) and those predicted by theoretical models ( 0.1 - 0.5% , hasegawa & herbst 1993 ; ruffle & herbst 2000 ) , and at least a factor 2 larger than the gaseous c@xmath0h@xmath0/h@xmath0o ratios obtained in observations of cometary comae ( 0.1 - 0.9% , brooke et al .
1996 ; weaver et al .
we speculate that these discrepancies might result from ( 1 ) the destruction of co@xmath0 by reaction with atomic hydrogen in shocks faster than @xmath5 30 km s@xmath36 ( predicted by charnley & kaufman , 2000 ) and/or ( 2 ) a greater efficiency for sputtering of c@xmath0h@xmath0 in slow shocks . would not necessarily require any fine tuning of the shock velocity .
in reality , any sight - line typically samples an ensemble of shocks with a _ range _ of shock velocities , and the constancy of the @xmath34 would simply indicate that the admixture of shock velocities varies little from one sight - line to another ( e.g. neufeld et al .
2006 ) . ] in either case , the gaseous c@xmath0h@xmath0/co@xmath0 ratios we observed may exceed the solid c@xmath0h@xmath0/co@xmath0 ratio , and the @xmath10(c@xmath0h@xmath0)@xmath35/@xmath10(h@xmath0o)@xmath35 ratio could be less than 0.02 . unfortunately , direct measurements of acetylene ice are not possible , the weak features expected from solid c@xmath0h@xmath0 being blended with much stronger features of co and h@xmath0o ( boudin et al . 1998 ) .
however , further observations of gaseous c@xmath0h@xmath0 at a higher signal - to - noise ratio would be very valuable as a probe of any variations in the @xmath34 ratio which might provide important clues to the shock physics .
this work is based on observations made with the _
spitzer space telescope _ , which is operated by the jet propulsion laboratory under a nasa contract .
d.a.n acknowledge funding from ltsa program nag-13114 to the johns hopkins university .
we are grateful to j.m .
torrelles for providing us with the nh@xmath25 maps .
we thank the referee for helpful comments . | the first map of interstellar acetylene ( c@xmath0h@xmath0 ) has been obtained with the infrared spectrograph onboard the _
spitzer space telescope_. a spectral line map of the @xmath1 vibration - rotation band at 13.7 @xmath2 m carried out toward the star - forming region cepheus a east , shows that the c@xmath0h@xmath0 emission peaks in a few localized clumps where gas - phase co@xmath0 emission was previously detected with _ spitzer_. the distribution of excitation temperatures derived from fits to the c@xmath0h@xmath0 line profiles ranges from 50 to 200 k , a range consistent with that derived for gaseous co@xmath0 suggesting that both molecules probe the same warm gas component .
the c@xmath0h@xmath0 molecules are excited via radiative pumping by 13.7 @xmath2 m continuum photons emanating from the hw2 protostellar region .
we derive column densities ranging from a few @xmath3 10@xmath4 to @xmath5 7 @xmath3 10@xmath6 @xmath7 , corresponding to c@xmath0h@xmath0 abundances of 1 @xmath3 10@xmath8 to 4 @xmath3 10@xmath9 with respect to h@xmath0 .
the spatial distribution of the c@xmath0h@xmath0 emission along with a roughly constant @xmath10(c@xmath0h@xmath0)/@xmath10(co@xmath0 ) strongly suggest an association with shock activity , most likely the result of the sputtering of acetylene in icy grain mantles . |
magic are two imaging atmospheric cherenkov telescopes for gamma - ray astronomy located at the _ observatorio del roque de los muchachos _ ( european northern observatory , la palma island , spain ) .
the production of a large amount of data during normal operation is inherent to the observational technique .
the storage and processing of these data is a technical challenge which the magic collaboration has solved by profitting from infrastructures like those developed for lhc experiments . + during the last years , the magic groups at ifae and uab ( barcelona ) and ucm ( madrid ) have set up , in collaboration with pic , the magic data center .
the facility became operational in february 2007 and , as of now , is equipped with the needed storage resources and computing capabilities to process the data from the first telescope and make them available to the magic collaboration .
however , as magic ii is expected to start generating data this year , we expect the data volume to be increased by a factor of 3 with respect to the present single telescope situation . in consequence
, we will increase the capabilities of the data center by providing it with the needed hardware and human resources to make it able to centralise the storage and analysis of the data .
the main goals of this extension are to allow fast massive ( re-)processings of all stored data and to support data analysis for all magic collaborators .
+ in what follows we will describe the present status of the data center and the foreseen upgrades required to deal with the data flow from the two - telescopes system .
this will lead us to provide additional services useful for the collaboration and also for the astrophysical community .
+ [ cols= " < , > , > , > , > " , ]
the magic telescopes have in their focal plane a camera segmented into a number _ c _ of different pixels ( each equipped with a photo - multiplier ) , whose signals are digitized by the daq system @xcite . currently , magic i is in service as a single telescope with a camera of 577 pixels .
magic ii is under commissioning , and has an improved camera with 1039 pixels for regular operation plus 42 additional pixels equiped with experimental high quantum efficiency photodetectors for test purposes @xcite@xcite .
+ for every trigger , the single pixel signal is sampled _ s _ times .
each sample is digitized with 12 bit precision and the resulting values stored in 2 byte fields .
the information is then saved into a raw data file .
the size of a magic raw event is given by _
+ 2byte__@xmath0s@xmath0c _ _ , where _ h _ is a fixed - size ( 4.5 kbyte ) header describing the event .
the event rate depends on the observation conditions and trigger configuration , and can range between 200 and 700 hz .
the average event rate during observation periods 67 - 73 ( may - dec 2008 ) was 350hz .
we will use this value to compute the data volume and storage needs summarized in table [ tab : volume ] . + raw event files
are processed using the magic standard analysis and reconstruction software ( mars ) @xcite .
the first step is a program dubbed _ callisto _ , which calibrates the cherenkov pulse s intensity and arrival time , producing a so - called calib data file .
this part of the analysis is the most cpu demanding , therefore calib data files are saved before further data processing .
the rest of the analysis chain consists of a set of executables taking as input the output of the previous program in the chain : _ star _ computes the parameters describing the cherenkov images of the individual telescope @xcite@xcite ; _ superstar _ merges the information of a given shower from the two telescopes ; and _ melibea _ computes the estimated energy , arrival direction and the so - called _ hadronness _ ( a parameter used for gamma / hadron discrimination @xcite ) . the output from _ star _ and _ melibea _ will be referred to as different steps reduced data files , and is also stored permanently .
+ estimations of the data volume at the different stages of the analysis chain , and for the different telescope configurations are also shown in table [ tab : volume ] , the different phases of the standard analysis , the name of the standard programs and the input and output file formats are summarized in table [ tab : analysis ] . + a diferent route is followed by the monte - carlo simulated events that are used for the estimation of the energy and _ hadronness_. in this case , instead of raw files , atmospheric particle showers are generated using corsika and then digested in two steps ( _ reflector _ and _ camera _ ) that finally produce data - like files that are calibrated and reduced with the same programs used for real data .
the parameters of the detector simulation are adapted to the telescope performance in different observation periods and configurations .
therefore , several versions of the monte - carlo library are provided at the data center .
presently , the simulated events are generated at the infn padova and udine , and the resulting files require 10tbyte of disk space . +
currently , the magic data center takes care of the following tasks : * data transfer from la palma to pic , via internet and tapes * data storage on tapes and disk at pic * data access at all data processing levels ( raw , calib , reduced ) for all magic collaborators * real - time , automatic analysis of the data , processed with magic standard software * reanalysis of all stored data in case of software updates and bugfixes * magic data base * software repository and bug tracker * storage of the data quality control files during normal data taking on the site , raw data are stored into disk , and later recorded to tape .
the tapes are then sent from la palma to pic via airmail , since currently there is not enough internet bandwidth between the island and europe to support the transfer of the files .
+ the computing system on the site also performs the so - called _ onsite _ @xcite analysis , by which raw data are processed , producing calib files and reduced files right after the data taking .
these files are indeed transferred to pic via internet , together with some log files generated by the subsystems of the telescope . + currently , all tapes received at pic are downloaded to a buffer disk and then written back to tape grouped by source and observation night in a single file ( an _ iso volume _ ) that can be mounted as an external unit in the file system .
this procedure is obsolete and recently has been found more optimal to store the files directly in tapes .
+ the data are organized in a data base and served to the collaboration through a user interface machine and a web site . until late 2008
these data were in a nfs file system , but currently , all data are being migrated to a new grid - based file system ( dcache ) that allows more transparent access ( in the sense of making no difference between tape and disk storage ) to any level of raw and processed data .
it is planned that in july 2009 the nfs will be finally dismantled and all applications and data access will be based on grid .
+ the reduction of the files that arrive via internet is triggered by scripts that run automatically every few minutes and notice when transfers have successfully finished . when this happens , batch jobs are generated and submitted to the grid with a specific configuration that ensures that they will run at pic ( the only computing element where mars is currently installed ) .
this set of scripts can also be used to massively reprocess all the data stored at pic in case that a bug is found in the software or some improvement in the analysis makes it worth .
massive operations of this kind have been performed twice in 2008 and once more in 2009 . for the last one
, we have estimated that we used 115 days of cpu time in two weeks .
this means that we have the capability to run _ star _ on one year of data in about 8 days .
it is worth to mention that this peak processing rate would have never been achieved with just the minimum number of cpu cores that the magic project is granted at pic according to its share .
in fact , we should always have access to at least 7 cores any time , but in periods of low usage from other experiments we have got up to 10 times these resources .
+ finally the data center also provides a `` concurrent version server '' ( _ cvs _ ) for the software development and the daily check , which generates a daily report on the data quality and the telescope stability .
in a near future we want to provide additional services to allow a more agile analysis by any magic collaborator and also easier access of anyone to published data . for this
we intend to : * extend the automatic data reduction up to _ melibea _ * provide resources and tools for high level analyses by any magic collaborator * open the magic public data to the whole scientific community by linking it to the european virtual observatory currently , the high level analysis , starting from melibea , is carried out by analysers that select a mc - gamma sample and a real data sample fitting well the observational conditions of the analyzed data .
these samples are used in the training of the multidimensional technique of selection of gamma - like events ( the random forest method @xcite ) .
we intend to automatize also this part of the analysis at pic in the near future , making the task of the analyzer simpler . + also
the computing power of pic can be more widely exploited by opening the job submission to the rest of the collaboration .
the already working roles of the grid scheme allow to assign priorities to the cpu farm users according to their duties , securing that the official data reduction is not delayed .
+ finally , we intend to establish a link with the european virtual observatory in order to share potentially interesting data for the astrophysical community . for this purpose a software that will translate root information in the widely used format in astronomy - fits
is currently being developed .
the magic data center based at pic is already providing quality services to the magic collaboration , exploiting when possible the extra resources that a grid - based infrastructure implies .
+ the success of the two year experience as official data center makes us push for the extension of the current facilities .
we hope this will improve the access to the data by the magic analyzers , and will make it more transparent for interested astrophysicists outside the collaboration .
+ 99 albert , j. et al .
instr . meth .
a 594 ( 2008 ) 407 orito , r. et al .
_ these proceedings _ borla , d. , goebel , f. et al .
_ these proceedings _ moralejo , a. et al . , _
these proceedings _ bretz , t. , wagner , r. , 2003 proc . of the 28th icrc ( tsukuba ) hillas , a. m. 1985 proc . of the 19th icrc ( la jolla )
aliu , e. et al . ,
. phys . 30 ( 2009 )
293 albert , j. et al .
meth . a 588
, 424 ( 2008 ) oya i. et al . ,
highlights of spanish astrophysics v cd - rom ( 2009 ) | the magic i telescope produces currently around 100tbyte of raw data per year that is calibrated and reduced on - site at the observatorio del roque de los muchachos ( la palma ) . since february 2007 most of the data have been stored and further processed in the port dinformaci cientfica ( pic ) , barcelona .
this facility , which supports the grid tier 1 center for lhc in spain , provides resources to give the entire magic collaboration access to the reduced telescope data .
it is expected that the data volume will increase by a factor 3 after the start - up of the second telescope , magic ii .
the project to improve the magic data center to meet these requirements is presented .
in addition , we discuss the production of high level data products that will allow a more flexible analysis and will contribute to the international network of astronomical data ( european virtual observatory ) . for this purpose
, we will have to develop a new software able to adapt the analysis process to different data taking conditions , such as different trigger configurations or mono / stereo telescope observations .
massive data processing , grid , european virtual observatory |
this review will start with a sketch of the kinematical - algebraic aspects of the overlap dirac operator in the vector - like context .
next comes a general discussion of numerical implementations of the overlap dirac operator .
section 2 is devoted to an alternative domain wall model .
this model is domain - wall like in the sense that an extra dimension is added and the computation of the light fermion propagator requires a single conjugate gradient procedure , albeit for a matrix representing fermions in five dimensions . on the other hand ,
the model is designed so that its output in exact arithmetic is the same as that of an iterative method implementing the overlap dirac operator directly .
the latter method requires a two - level nested conjugate gradient procedure . in section 3
rigorous results on spectral properties of our model are presented and in section 4 these results are compared to true domain wall fermions .
the main conclusion is that nested procedures typically are more efficient than implementations based on domain - walls .
this counter - intuitive conclusion is explained by the condition number of the domain wall problem being the product of the condition number of the four dimensional problem by the inverse bare quark mass .
the latter two factors govern individually the two nested cycles in direct implementations . in section 5
it is shown that one can eliminate the requirement of linearly growing memory consumption for increasing accuracy at the cost of a factor of two in operations . in practice ,
the factor of two is often not felt because the implementation is memory bound .
section 6 contains some final comments .
a large part of this talk is about work done in collaboration with rajamani narayanan @xcite .
the overlap formulation @xcite of vector - like gauge theories on the lattice preserves chiral symmetries exactly , a property thought to be unattainable for many years .
if one adds to the ginsparg wilson relation , as originally formulated in 1982 @xcite , a requirement of @xmath0-hermiticity , the combination is equivalent to the overlap at the algebraic level .
the set - up has been of interest to mathematicians much earlier as a generalization of the concept of angle between two straight intersecting lines in a plane .
the plane is generalized to a vector space over the complex numbers and the two lines to two subspaces of the vector space @xcite . in our application
the dimension of the vector space will always be even , but the subspaces can have unequal dimensions - when they do , the angle concept looses its meaning .
the kato setup is also meaningful for infinite dimensional hilbert spaces and has other applications to physics @xcite .
the subspaces are defined by projectors onto them , in our case the projectors are hermitian and replaced by linearly related reflections , @xmath1 and @xmath2 . @xmath3 and @xmath4 .
the @xmath5 eigenspaces of @xmath1 and @xmath2 are the subspaces in question .
they are spanned by orthonormal sets denoted by @xmath6 and @xmath7 with @xmath8 .
the information about the relative positioning of the subspaces is contained in the overlap matrix @xmath9 : @xmath10 the coarsest measure of relative orientation is obviously @xmath11 .
the main identity @xcite is @xmath12 in our case @xmath13 and @xmath14 , where @xmath15 and @xmath16 is any lattice version of the dirac operator describing fermions with negative mass of order @xmath17 where @xmath18 is the lattice spacing .
@xmath2 can be thought of as describing dirac fermions of positive infinite mass .
@xmath16 is restricted by the requirement that @xmath19 be hermitian . the simplest choice , which minimizes operations in the computation of the action of @xmath16 on a vector , is to pick @xmath16 as the wilson lattice dirac operator .
it is possible that when all practical aspects of a simulation are taken into account a seemingly more complicated choice would pay better off .
@xmath19 depends on the background gauge field represented by a collection of unitary link matrices @xmath20 .
@xmath19 has the following properties : * @xmath19 is too large to be stored in the computer memory in its entirety , but it is sparse , so its action on vectors can be implemented . *
the spectrum of @xmath19 is bounded by a finite bound that does not depend on the gauge background . * if all products of gauge matrices on links around plaquettes , @xmath21 , obey @xmath22 , where @xmath23 is a small positive number independent of the gauge background , the spectrum of @xmath24 is also bounded by a finite number , independently of the gauge - field .
the calculation of the action of @xmath1 on a vector must use sparse matrix techniques .
the boundedness of the spectrum means that the sign function needs to be approximated accurately only in two finite segments symmetrical about zero , which contain the entire spectrum of @xmath19 , for any gauge background allowed by the pure gauge action
. the main strain on the approximation occurs around zero , where the gap turns out in practice to be very small .
this is where essentially all the numerical cost goes .
there are two main approaches to the approximate implementation of the sign function : one is the direct approach ( overlap ) , and the other is indirect ( domain wall fermions )
. in the direct approach one looks for a rational approximation for the sign function in the range defined by the bounds on the spectrum of @xmath19 .
the rational approximation is written as a sum of pole terms .
the crucial point is that the action of each pole term on a vector need not be calculated separately : rather , the action of all terms can be calculated simultaneously , in one single pass through the conjugate gradient algorithm @xcite .
the approximation gets more accurate when the number of pole terms is increased .
settling for a certain number of terms , @xmath25 , one obtains an approximation for @xmath1 , @xmath26 the `` masses '' @xmath27 have to be non - negative , but the weights @xmath28 can take either sign . in practice we need the action of @xmath29 on a vector , so the action of @xmath30 is needed many times for one evaluation .
we end up with two levels of nested conjugate gradient algorithms . in the context of domain wall fermions
an approximation for the sign function is not constructed directly .
rather , one invents a larger problem , defined by a matrix @xmath31 .
@xmath31 is determined by @xmath19 but its dimensions are @xmath25-times larger .
@xmath31 is still sparse .
one then arranges that in a given subspace of the larger space @xmath31 acts on one has @xmath32 above , @xmath33 and @xmath34 can differ from eq .
( 2 ) by terms that disappear in the continuum limit .
the exact form is a matter of convenience .
one ends up needing to do only one inversion , using a single conjugate gradient algorithm ( cg ) . in the most common applications of domain wall fermions
one uses a construction of @xmath31 due to kaplan @xcite . historically , kaplan s formulation came first , and was a prime motivator for subsequent developments .
the main difference between the two methods is that one has a nested cg in one and a single cg in the other , but employing a larger matrix .
the objective of my work with narayanan @xcite was to invent a model containing a version of @xmath31 which is close to kaplan s domain wall fermions , but also to the rational approximation so that a comparison of numerical costs may be carried out using rigorous methods .
we wanted the output of either method to be the same in exact arithmetic , so it would be only the ways of doing the calculation that differ .
one way is similar to domain wall fermions and the other is direct .
let @xmath35 be the dirac field describing a light quark .
we wish to end up with an effective action for @xmath35 given by : @xmath36\psi .\ ] ] to be specific , we choose @xmath37 with @xmath38 this rational approximation can be easily replaced by others .
we now add @xmath39 extra fields , @xmath40 and @xmath41 , pack them together with the light field into a combined field , @xmath42 . the total action is @xmath43 introduce @xmath44 and @xmath45 .
the matrix @xmath31 is given by : @xmath46 our goal is attained by the model because of the identity : @xmath47 \times \nonumber\\ \int d\bar\psi d\psi e^{-\bar\psi \left ( { { 1+\mu}\over 2 } \gamma_5 + { { 1-\mu}\over { 2n}}\sum_{s=1}^n { 1\over { c_s^2 h_w + { { s_s^2}\over h_w } } } \right ) \psi } \end{aligned}\ ] ] the prefactor can be canceled by adding pseudofermions , which will be decoupled in the @xmath48 index .
let us now roughly estimate computational costs . in the direct approach the number of inner iterations is approximately given by the condition number of @xmath19 , @xmath49 ( which is the square root of the condition number of @xmath50 ) .
the number of outer iteration goes as the condition number of @xmath33 which is roughly @xmath51 .
thus , the total number of @xmath52 actions is roughly given by @xmath53 . in the domain wall version
the number of @xmath54 operations is governed by @xmath55 .
every @xmath54 operation counts roughly as @xmath39 @xmath19 operations .
we need to estimate @xmath55 in terms of @xmath56 and @xmath57 . to find @xmath55
we need the maximal and minimal eigenvalues of @xmath58 .
the basic trick for finding @xmath59 and @xmath60 is the derivation of an exact formula for @xmath61 .
@xmath62 \times \nonumber\\ \det \left [ { { 1+\mu}\over 2}\gamma_5 + z + { { 1-\mu}\over 2 } f_n ( h_w , z ) \right ] \end{aligned}\ ] ] here , @xmath63 eigenvalues of @xmath31 are roots of the equation @xmath64 .
all the roots come from roots of the last factor .
( roots of the factors in the product over @xmath48 are canceled by poles in the last factor .
so , the spectrum of @xmath31 is determined by the last factor . )
we write @xmath65 where , @xmath66 the point is that one has explicit formulae for @xmath67 , so long @xmath18 and @xmath68 are real . if @xmath69 we have @xmath70 if @xmath71 , @xmath72 these formulae make the @xmath25-dependence explicit .
let me describe the main idea for deriving bounds on the eigenvalues of @xmath31 .
we are looking for zero modes of : @xmath73 conditions on @xmath74 , depending on @xmath57 and spectral properties of @xmath19 , limit the eigenvalues of @xmath75 to some range .
suppose @xmath76 is a zero mode of @xmath77 .
it can exist only if @xmath78 is in the range of eigenvalue of @xmath79 .
but , @xmath80 is bounded from above and below for @xmath81 and for @xmath82 .
so , it is possible to exclude vicinities of @xmath82 and @xmath83 . in this way we obtain rigorous bounds :
@xmath84 and @xmath85 @xmath86 is a function of the spectrum of @xmath19 and of @xmath57 .
it is well defined and calculable but clumsy to write down . for ranges of practical interest @xmath87 .
this leads us to : @xmath88 where the last factor is close to unity in practice .
the meaning of the result is quite obvious : we need to overcome both the lattice artifact of having at times almost zero modes for @xmath19 and the physical effect of small quark mass ( @xmath57 , when close to zero , gives the bare lattice quark mass ) .
clearly our conclusion speaks in favor of the direct approach , if we assume the bounds to be typically saturated ( the action of @xmath31 is @xmath39 more expensive than that of @xmath19 ) .
but , before jumping to this conclusion we should try to decide how good our bounds are .
it turns out they are quite good .
the bound on the @xmath59 is typically saturated . as far
as @xmath89 goes we can prove @xmath90 ^ 2 \right ) \end{aligned}\ ] ] one can not do better than @xmath91 either . for @xmath57 small enough and @xmath25 large enough @xmath92
thus either a very small eigenvalue of @xmath93 or a very small mass are guaranteed to make the smallest eigenvalue of @xmath58 very small too .
the lower bound in eq . ( 18 ) , if optimal , indicates that it is possible for the smallest eigenvalue of @xmath58 to be as small as the product of these two small numbers .
for true domain wall fermions with action @xmath94 we can prove only one kind of bound : @xmath95 ^ 2 \nonumber\\ + ( 1+\mu)^2 \left ( { 1\over n } -{1\over { n^2 } } \right ) \end{aligned}\ ] ] an approximate analysis yields @xmath96 upper bounds on @xmath97 are of order 10 in practice .
we conclude that the matrices describing the version of domain wall fermions used in large scale simulations have conditions numbers that behave similarly to the condition number of our model .
= 2.5 in in a modest numerical experiment on a two - dimensional @xmath98 model we compared domain wall fermions using our alternative model , true domain wall fermions and the direct overlap approach .
the pure gauge action was of the single plaquette type , with a lattice coupling @xmath99 and the lattice size was @xmath100 .
we performed the calculations needed to obtain the condensate @xmath101 .
we required the norm of the residual to go down to @xmath102 and counted operations of @xmath19 on a vector .
the value of @xmath25 used was @xmath103 .
we did not use preconditioning in any of the three methods .
results were obtained using 20 gauge field configurations .
figure 1 shows the number of @xmath19-on - vector operations as a function of quark mass .
the comparison between the direct overlap and the alternative method is simple because , by design , the numbers one would get in exact arithmetic are the same in either method . comparing to true domain wall fermions is more difficult because one needs to match parameters to get similar physics , and this is ambiguous .
for example we set the wilson mass parameter in @xmath19 to @xmath104 in all three cases and used the same mass parameter @xmath57 although this certainly is not correct for very small @xmath57 .
we found that the best was to use the direct overlap approach , but this experiment was very limited and one should not immediately draw conclusions about numerical qcd .
however , there is enough evidence by now to scrutinize seriously the question whether it is worth investing large scale computational resources into true domain wall fermions given that there is a direct overlap alternative . in particular
we have employed no preconditioning and ignored many other options of optimization .
theoretically , one is more comfortable with overlap fermions , since they differ from other four dimensional fermions just by a more complicated action . if the domain wall approach were more effective computationally one could have the best of the two worlds by using our alternative domain wall fermions .
whichever the case may be , it is hard to argue in favor of true domain wall fermions .
until now we concerned ourselves only with counting operations .
but , as is well known , this is only one aspect of a computation .
memory requirements and access patterns matter as much , or more , depending on architecture . until now all methods required a factor of @xmath25 times more memory than on ordinary wilson fermion code would . in the direct approach
there is an option to trade a factor of order 2 in operations for reducing the memory needed to that of ordinary wilson fermions .
basically , this is possible because the heart of the code is a conjugate gradient procedure , and only a certain linear combination of the @xmath25 vectors that are acted on is needed .
the conjugate gradient procedure is closely related to the lanczos scheme for bringing a hermitian matrix to tri - diagonal form .
it is well known that in the lanczos scheme , if one wishes additional information ( for example one needs an eigenvector ) one can avoid large memory consumption by going through the algorithm twice , once to collect the coefficients and a second time to accumulate the needed data .
the same works here @xcite . in the multishift scheme @xcite
one iterates over an index @xmath105 updating @xmath25 vectors @xmath106 , starting from an input vector @xmath68 . at each iteration
these vectors are determined by a few @xmath48-dependent scalars and by three @xmath48-independent vectors that make up the core conjugate algorithm , @xmath107 . among them , @xmath108 is the residual .
when the iteration is stopped we have @xmath25 @xmath109 vectors and they are summed into @xmath110 , our approximation to the vector @xmath111 @xmath112 actually , the iteration must be effectively stopped after a different number of steps for each @xmath48 , since for higher values of @xmath48 the `` mass '' is larger and , as a result , the convergence much faster .
clearly , @xmath110 is made out of the basic krylov vectors generated in the core conjugate gradient ; the single reason we need to store @xmath25 vectors is that the components of @xmath110 in the krylov basis at step @xmath105 in the iteration are not yet known because they depend also on steps @xmath113 .
the idea is now to use a first pass to calculate the needed conjugate gradient scalars which are @xmath105 dependent but not @xmath48-dependent . using these scalars we can now compute in an iteration storing scalars only the extra @xmath48 and @xmath105-dependent scalars needed for implementing the multishift .
it is possible now to also compute a set of @xmath48-independent scalars @xmath114 which have the property that @xmath115 becomes at the last iteration the desired approximation for @xmath110 .
but , we need the vectors @xmath116 again so we need to run the basic conjugate gradient again . hence , a second pass is required , but one needs to store only four large vectors for any @xmath25 .
a more detailed description of the algorithm can be found in @xcite .
the surprise was that with a code written in a higher level language the two pass version can actually run faster than the single pass version , by about 30 percent .
in the two pass version both the operations count and the memory usage are independent on @xmath25 !
the speed up is certainly a surprise and must be strongly machine dependent .
but , it turns out , that one gets the same amount of speed - up on an sgi o2000 @xcite as on a pentium iii pc @xcite .
let us summarize roughly the situation we were looking at when we began to work on reference @xcite : although an approach based on the overlap dirac operator looked theoretically cleaner , true domain wall fermions were more attractive numerically .
our analysis has led us to the conclusion that there is no evidence that true domain wall fermions have even a numerical advantage . in all cases we looked at ,
one faces a problem related to almost zero modes of @xmath19 .
this requires large numbers of extra fields in order to preserve chirality .
it also affects adversely the condition numbers .
whichever method we use , the worst case condition numbers are a product of the inverses of two main scale ratios : the first is the scale of the small eigenvalues of @xmath50 divided by an upper bound of the order of 5 - 10 in lattice units . the second scale ratio is the lattice physical quark mass squared divided by a number of order unity .
each small scale ratio slows down inversion independently and the effect compounds in the worst case .
thus , as far as we can see , at the numerical level , there are no _ a priori _ advantages to choosing true domain wall fermions over overlap fermions in the context of qcd . in both formulations
one faces similar numerical obstacles , and the overlap , to say the least , does not fare any worse than domain wall fermions . at the analytical level
we are convinced that an approach based on the overlap ( or any other efficient replacement of the overlap dirac operator that might be found in the future ) is superior at presently attainable gauge couplings in numerical qcd .
perturbation theory is more transparent to interpret and technically less complex in the overlap version .
the chirality violating effects associated with the number of extra fields are much more explicit and therefore their impact should be easier to trace through .
algorithmically we have been looking only under the lamp post and we are far from having exhausted the options there .
nevertheless , i feel that true domain wall fermions are an inefficient way to incorporate the new ideas about lattice chirality into practical qcd simulations .
for other recent work on similar issues the reader is referred to @xcite .
i would like to urge you to use your imagination : there must exists much better ways than the ones we have tried until now .
this research was supported in part by doe grant de - fg02 - 96er40949 and by a guggenheim fellowship .
i would like to thank l. baulieu and the entire group at lpthe for their hospitality and support .
9 r. narayanan , h. neuberger , phys .
d62 , 074504 ( 2000 ) d. b. kaplan , phys .
b288 , 342 ( 1992 ) ; r. narayanan and h. neuberger , phys .
b302 , 62 ( 1993 ) ; nucl .
b443 , 305 ( 1995 ) ; h. neuberger , phys .
81 , 4060 ( 1998 ) ; phys .
b417 , 141 ( 1997 ) ; phys .
b427 , 125 ( 1998 ) ; phys .
d57 , 5417 ( 1998 ) ; phys .
d60 , 065006 ( 1999 ) ; phys .
d61 , 085015 ( 2000 ) p. ginsparg , k. wilson , phys . rev .
d25 , 2649 ( 1982 ) .
t. kato , `` perturbation theory for linear operators '' , springer - verlag , berlin , 1984 .
j. avron , r. seiler , b. simon , comm .
phys 159 399 ( 1994 ) .
a. frommer , s. gsken , t. lippert , b. nckel , and k. schilling , int .
c6 , 627 ( 1995 ) ; b. jergerlehner , hep - lat/9612014 . h. neuberger , int . j. mod .
c10 , 1051 ( 1999 ) .
j. kiskis and r. narayanan , phys .
d64 , 117502 ( 2001 ) . and private communication .
r. edwards , u. heller , phys .
d63 094505 ( 2001 ) . | an alternative to commonly used domain wall fermions is presented . some rigorous bounds on the condition number of the associated linear problem
are derived . on the basis of these bounds and some experimentation
it is argued that domain wall fermions will in general be associated with a condition number that is of the same order of magnitude as the _ product _ of the condition number of the linear problem in the physical dimensions by the inverse bare quark mass .
thus , the computational cost of implementing true domain wall fermions using a single conjugate gradient algorithm is of the same order of magnitude as that of implementing the overlap dirac operator directly using two nested conjugate gradient algorithms . at a cost of about a factor of two in operation
count it is possible to make the memory usage of direct implementations of the overlap dirac operator independent of the accuracy of the approximation to the sign function and of the same order as that of standard wilson fermions . |
understanding the fundamental properties of planetary nebulae ( pne ) requires knowledge of their distances . for planetary nebulae , because of the absence of a standard quantity common to all , there is no well - calibrated standard distance scale and distances estimated from different independent methods often disagree with each other .
one way to determine the distance of a planetary nebula is to measure its angular expansion during a period of time and combine this with its doppler expansion velocity .
this method has been used for both vla ( very large array ) radio maps @xcite , and optical images from the _ hubble space telescope _ ( _ hst _ ) @xcite .
( hereinafter , `` bd+30 '' ) is a rapidly evolving , chemically inhomogeneous nebula @xcite .
it is one of the few nebulae with detected x - ray flux from its bubble of shocked stellar wind @xcite .
this justifies extensive observations of this object . to understand the rapid chemical evolution of bd+30
, we need to know the mass and luminosity of its central star , and this requires a reasonably accurate distance .
the earliest expansion distance of bd+30 was obtained by @xcite from vla maps , who gave a result of @xmath7 kpc .
@xcite used different two - epoch vla observations , resulting in a distance of . @xcite
combined the observations used by @xcite with a new set of data obtained in 1993 , and derived the most accurate measurement at that time of @xmath8 kpc . to obtain a more accurate determination of the distance , and details of its expansion
, we used the two - epoch optical observations of _ wide field planetary camera _ ( _ wfpc2 _ ) on board the _ hst _ to measure the expansion of this nebula .
the total expansion of bd+30 was estimated to be about 15 mas between the first and second epochs , which is about 1/3 of the _ planetary camera _ ( _ pc _ ) 00455 pixel .
this tiny displacement in diffuse objects can be measured because of the high resolution of _
images @xcite , and we can do even better for a number of diffuse knots sharing a common proper motion .
much evidence shows that bd+30 is not actually a spherical object @xcite .
we thus use an ellipsoidal kinematic model of bd+30 to combine the angular expansion with the radial expansion velocities from spectroscopic observations to find the distance and kinematic age of this nebula .
_ hst _ narrow - band images from our go programs 5403 and 8116 were used for this study ( table [ observtable ] ) ; the observations were separated by 5.663 years . for both epochs , images through f656n ( h@xmath0 ) and f658n ( [ ] ) narrow - band filters were obtained . in the first epoch
there were two images for each band with the same pointing but different exposure times .
the images with longer exposure time were saturated in places . for the second epoch
, there were three images for each band , with the same exposure time and orientation , but slightly shifted by fractional pixels with respect to each other for the purpose of drizzling . since the orientation differs for the two epochs , the images had to be aligned to a common center and orientation
the data reduction includes cosmic ray rejection , drizzling , and image alignment .
other adjustments or corrections like geometric distortion correction , saturated pixel masking , were also done during the three main steps . because the 2nd epoch h@xmath0 and [ ] images were dithered cosmic rays can not be removed by the commonly used iraf task _
crrej_. however , @xcite and @xcite introduced a method to remove cosmic rays from a group of dithered _ hst _ images , as well as an iraf package _ drizzle _ for this purpose .
we first used iraf task _ cross_driz _ , _ shiftfind _ , and _ imshift _ to find the relative displacement between images , and align the images .
then , we used _
crrej _ to combine the aligned images to produce cosmic - ray - free images . finally , _
driz_cr _ was used to create cosmic ray masks , which were used in the drizzling step . using the iraf package _ drizzle _ , three 2nd epoch h@xmath0 images and three 2nd epoch [ ] images with _
pc _ image resolution 800@xmath9800 were combined into images with resolution 1600@xmath91600 , respectively . to compare images from the two epochs ,
the old ( the 1st epoch ) images were also mapped into the same 1600@xmath91600 grid by the _ drizzle _ package , and cosmic rays were removed by the same strategy as for new ( the 2nd epoch ) images .
it is easy to tell that some features are sharper in the new drizzled images than in the old images ( which were merely mapped onto finer grids ) .
the drizzled h@xmath0 image is shown in fig .
[ drizzled_halpha ] . to perform the comparison between new images and the old ones , images had to be aligned in displacement and position angle .
the relative rotational angle and shift between two epochs of images were found by iraf task _
cross_driz _ , _ rotfind _ , and _ shiftfind _ , iteratively .
the most ideal and simplest situation is that the expansion is isotropic , so that it looks symmetrical .
we took this as the 0th order approximation , and did the alignment by trail - and - error strategy with the following measurements to make the measured average expansion along radial lines as symmetrical as possible .
by blinking the aligned images taken at two different epochs back and forth , the expansion can be seen clearly by eye .
the expansion in 1600@xmath91600 images is estimated about 0.5 to 1 pixel ( or 0.25 to 0.5 pc pixels ) .
the actual measurements were performed using several different independent methods .
basically , there are two ways to measure the expansion ; one is magnification , which is whole - image - based , the other is to measure the shifts of individual features . if the expansion is spherically symmetric , the new image should be a magnified version of the old image .
this method can be applied to the whole image , or along sectors in particular directions from the image center . because the old ( first epoch ) images have many saturated pixels ,
while the new ( second epoch ) ones do not , we reduced the new images to fit the old images , which avoids any manipulation of the saturated - pixel masks .
the reduction factor was determined by minimization of the square of the difference image rather than by cross - correlation , which we found less sensitive .
we estimate the uncertainty at 25% for the global magnification factor and for the magnification factors along radial lines .
however , since the expansion of the nebula may be asymmetrical , the magnification method can not provide the details of the angular expansion .
thus we made measurements of individual features and took the results of the magnification method as a consistency check . compared to other pne such as ngc 6543 , bd+30 is more diffuse , so it is not possible to measure small regions of only a few pixels in size . to determine the best size of regions for shift measurements , we tested the effects of square size on some randomly selected features by plotting the curve of measured shifts against the size of the region .
we found that when the square size was around 20 pixels ( i. e. , 10 pc pixels ) , shifts were the most insensitive to the size of the region .
we therefore adopted 20 @xmath9 20 pixel squares for the measurements .
the basic compare - and - fit strategy used here is just like what is used to produce a cross - correlated image .
first , move one image by , for example , pixels in the @xmath10 direction and pixels in the @xmath11 direction . then compare and quantify the difference by summing up the squares of differences of corresponding pixels as in a least - squares fit .
this gives the value of the point in the least - squares image .
finally , find the minimum of the least - squares image by gaussian 2-d fitting .
this position gives the relative shift between two features .
care must be exercised in the treatment of the edges of the squares , since almost all regions cut from the original images have non - zero edges ( and the edges of saturated pixel masks are also sharp discontinuities ) .
tapering is not satisfactory , because the regions are very small , as are the shifts ; tapering the edges will drag the measured shifts to zero .
the method we employed was to treat the 20 pixel square , combined with the saturated pixel mask within the square region , as a window .
we first moved the second epoch image , which contained no saturated pixels , then looked at both images through that window , and compared the visible parts by least squares . to simplify the process , idl routines were developed which can select small regions from images interactively , and calculate the shifts automatically .
we used the magnification method at two levels , one for whole image , another along every radial line at 10@xmath4 intervals from 0@xmath4 to 360@xmath4 .
numerical fitting for the whole image gave magnification factors of 0.99384 for h@xmath0 and 0.99361 for [ ] .
the results from radial line magnification are shown in fig .
[ magniflines ] . as mentioned above ,
one criterion of alignment was to make the angular distribution of magnification as symmetrical as possible .
even so , we can still notice that the measured expansion shows some structure .
the expansion is greatest near the openings in the northeastern and southwestern quadrants .
while the distribution of walls and openings is irregular , the smallest expansion seems associated with regions where the shell is brightest .
the direction of greatest expansion is not along the longest axis of the nebula s image ; this may be because the nebula is a tri - axial ellipsoid , with none of the axes aligned with the line of sight , so the expansion might not be aligned along the longest axis of the image .
this might also explain why some features near the center were measured to have inward displacements .
the shifts of nearly 200 features in the h@xmath0 images and the [ ] images are shown in fig .
[ featureh ] and fig .
[ featuren ] , respectively , which illustrates the details of expansion .
again , except for a few features , the results from the two independent wave - bands show a high degree of consistency ( fig .
[ compare ] ) .
the agreement between h@xmath0 and [ ] gives us confidence in the results .
the discrepancies between the two wave - bands may partly be due to slight misalignment , so that the same coordinates in different images might refer to slightly different regions .
in addition , since the ratio of [ ] to h@xmath0 emission varies from point to point within the nebula , there may be real differences in the shifts of features seen in the two wave - bands .
the expansions of the major axis and minor axis of the optical shell are summarized in table [ exptable ] .
the kinematic age of the nebula , defined as @xmath6 , varies somewhat with the method of measuring the expansion as well as from region to region .
a good average value is 800 years .
the magnification factor derived for the whole image gives kinematic ages of 920 and 880 years for h@xmath0 and [ n ii ] , respectively .
the magnification factors along sectors give somewhat smaller ages , and the ages along the minor axis tend to be less than along the major axis .
finally , expansions from feature shifts give the smallest kinematic ages , 740 - 800 years , and are again smallest along the minor axis .
as discussed below , we are particularly interested in the expansion along the position of a spectrographic slit at position angle 99@xmath4 ; the kinematic age along this direction is about 600 years .
if we want to derive the distance of the nebula , some kind of kinematic model implicit or explicit must be employed to relate the tangential expansion rate to the expansion velocity along the line of sight .
@xcite recently discussed the kinematics of bd+30 based on ground - based optical emission - line spectra .
they measured the spectra along two slits roughly aligned to the east - west and north - south directions through the center , which , approximately , are the major and minor axes of the nebula . for the modeling of bd+30 , they used a tilted ellipsoidal shell , with a high velocity expanding h@xmath12 ring in the equatorial plane .
furthermore , they found that there is a difference between the low ionization and high ionization regions of this nebula . compared with low ionization regions ,
the high ionization regions are smaller , but have much higher velocities .
they found the expansion velocity along the line of sight to the central star to be @xmath13 km s@xmath3 for [ ] profiles and @xmath14 km s@xmath3 for [ ] profiles .
because the nebula is compact , ground - based long - slit spectra do not have the optimum degree of spatial resolution .
as part of our hst program , we obtained stis spectra with the e230h echelle grating using a @xmath15 x @xmath16 slit .
the grating was set to include the strong ] multiplet @xmath12324.21 , 2325.40 , 2326.11 , 2327.64 and 2328.84 , as well as the [ ] doublet @xmath12470.97 , 2471.09 .
the echelle mode of stis is available only with the ultraviolet mama detector , so substantial dust extinction was unavoidable .
two position angles were observed , 25@xmath4 and 99@xmath4 .
the 25@xmath4 orientation suffers badly from extinction , especially to the north , so we we rely primarily on the 99@xmath4 p.a .
the slit was positioned so that it passed @xmath17 north of the central star to avoid contamination by the stellar continuum .
the location of the 99@xmath4 slit on the h@xmath0 image is shown in fig .
[ slitfit ] . to increase the s / n ratio ,
the spectrum was shifted by 1.53 and added to itself to superimpose c ii ] 2327.64 and 2326.11 , the two strongest components .
the results are shown in fig .
[ spectrum ] .
panel a is the 25@xmath4 orientation and b1 and b2 the 99@xmath4 slit . in this figure
, the top of the slit is to the east .
the [ o ii ] line appears identical to c ii ] , but has lower s / n , so we do not discuss it further .
we see that the line has well - defined red and blue components , which yield an accurate expansion velocity .
if the nebula were a uniformly expanding ellipsoid , the echelle profile would be an ellipse tilted , if the ellipsoid major axis were inclined to the line of sight . while our profile is roughly elliptical , there are clearly local irregularities in the expansion .
panel ( b1 ) shows an ellipse which fits the data reasonably well .
the vertical extent of this ellipse is 494 and the horizontal splitting ( at the center ) is @xmath236.25 km s@xmath3 .
the ellipse is tilted to the right by 5@xmath4 from vertical .
while the splitting seems to be well determined , the tilt of the ellipse is not secure . but basically , it is a tilted ellipse , which indicates that a tilted ellipsoidal shell is an appropriate model of the nebula .
the lower edge is fainter due to local dust extinction @xcite . since c ii , like n ii , is a lower stage of ionization , we were surprised that the expansion velocity of this line was close to the result @xcite found , not for [ n ii ] , but for the high ionization [ ] line
. perhaps our higher spatial resolution gives a more accurate velocity , while the lower ground - based resolution fills in the profile center , lowering the average velocity .
another consideration is that , although the c@xmath18 ions will avoid the highly ionized central regions , the excitation level of this uv line is high , so the emission will be biased toward the hottest parts of the nebula .
because the expansion velocity was only measured along the stis slit , we made measurements of the astrometric shifts of features along that cut .
according to our kinematic model , the tangential expansion is along radial directions , and is linearly proportional to the distance from the center . under these conditions ,
projections of expansion velocities on any axis are proportional to the corresponding projections of distances on that axis .
so for point of the image where the center is at , its expansion velocity satisfy and , and it does not matter which direction you specify as @xmath10 direction .
therefore , we extracted the components of the shifts along the slit for each feature along the cut , and plotting them against the distances to the center of the slit , fitted them to a straight line , as shown in fig .
[ slitfit ] .
the least - squares line goes through the center , which is a good demonstration of symmetry .
obviously , the error of the central point will be large , since its shift should be perpendicular to the slit .
a fit excluding the center three points , which are likely to have large errors , gave almost the same line . in this way
, the expansion rate at the edge of the nebula , which corresponds to the upper and lower edge of the spectral ellipse , was measured to be at @xmath19 away from the center . if the nebula were a uniformly expanding spherical shell , then the above tangential expansion , combined with the 36.25 km s@xmath3 velocity along the same cut , implies a distance of 1.80 kpc .
there is evidence , however , that the nebula is not spherical . in the next section
, we will attempt to develop a better kinematic model .
bd+30 is optically thick to ionizing radiation
it is surrounded by a neutral halo @xcite but the photoionized shell is also geometrically thin ( perhaps because of the pressure of the @xmath20 k x - ray emitting gas which fills the interior of the shell ) . under these conditions ,
the variation in the surface brightness across the nebula contains information about the three dimensional structure of the nebular shell .
since it would be difficult to correct for the internal extinction due to dust which affects the optical images , we have used the 5 and 15 ghz vla radio maps described in @xcite for this analysis .
a similar approach was previously used by @xcite , who modeled the apparent ew elongation and surface brightness variations of bd+30 in terms of an inclined prolate ellipsoid tilted from our line of sight along the ew direction , consistent with a limited kinematic information available at that time .
the relevant equations are developed in the appendix .
the constancy of flux in pie - shaped sectors radiating from the central star supports the idea of complete absorption of the ionizing radiation .
the ratio of the central surface brightness to the total flux indicates that the nebula is elongated along the line - of - sight : the central intensity is substantially less than would be the case for a sphere .
unfortunately , the angular expansion vanishes near the center of the nebula ; what we need is information on the shape of nebula where we have measured the angular expansion .
we can , however , find the shape of the shell along a chosen wedge from the surface brightness distribution ( eq .
we integrated the surface brightness in 25@xmath4 sectors chosen to overlap the region cut by our 99@xmath4 stis slit .
the results are shown in fig .
[ shell_sect ] .
the values near the central star are very noisy due to the low signal at the wedge apex .
also , near the edge of the nebula , the finite shell thickness invalidates the assumptions of the method .
but overall , the shell seems to be well fit by an ellipse with an axial ratio of 1.5 to 1 along our line - of - sight , as indicated by the dotted line in fig .
[ shell_sect ] .
this fit also closely matches the shell distance of 37 in the direction of the central star the triangle in fig .
[ shell_sect ] found from the central intensity ( eq .
a1 ) . in obtaining the shape from eq .
we made the assumption that the front and back surfaces of the shell contribute equally to the surface brightness .
but if we look at the echelle profile , we see that there is a great asymmetry between the front ( blue - shifted ) and back ( red - shifted ) branches .
we might attribute this to dust extinction , and indeed we feel that the relative weakness of the lower ( west ) part of the profile is likely due to dust . but
this can not explain the relative brightness of the upper , red - shifted side of the line , because this radiation comes from the _ back _ side of the nebula : this radiation must suffer at least as much extinction as the weaker blue side .
a slit placed across a uniformly expanding shell in the form of a tilted ellipsoid will result in a spectral line in the form of a tilted ellipse .
( the tilt angle of the major axis in a plot of the line profile will vary depending upon the relative scales of chosen for the x - axis and y - axis .
the ratio of the x displacement of the top of the ellipse to the width of the ellipse on the x - axis is , however , an invariant . ) because the ellipse that best fits our echelle profile seems tilted , we considered the appearance of a tilted ellipsoidal shell .
equation a7 gives the front and back surface brightnesses , which we found , for the parameters relevant here , to differ by up to a factor of two .
we can sum the front and back contributions and apply eq .
a3 to the result , to simulate our analysis of the vla data assuming symmetry .
we find that an ellipsoid of axial ratio 1.56 , tilted by 10@xmath4 , closely resembles an untilted ellipsoid of axial ratio 1.5 .
( such a solution is shown on fig .
[ shell_sect ] as a dashed line . ) since the top of our echelle spectrum seems tilted redward , this would imply an ellipsoid with the near side tilted down ( westward ) .
unfortunately , with such a tilt , it is the front ( approaching , blue - shifted ) side which is brighter at the top ( east ) end of the slit , just the opposite of what is observed . in an attempt to clarify this situation
, we examined the spectra presented by @xcite , who reproduce two spectra taken with an east - west slit ( close to the orientation of our 99@xmath4 slit ) , one in the [ o iii ] @xmath15007 line and the other in the relatively weak [ n ii ] @xmath15754 line .
the [ o iii ] profile is clearly tilted with the east side blueward , _ opposite the apparent tilt of our c ii ] line .
_ the [ n ii ] profile also seems to be tilted with the east side blueward , but determination of the tilt is difficult in this case because the parts of the profile corresponding to high velocities are very faint compared to the bright parts at lower velocities .
we find that if we apply the same analysis as above to the stis profile , this time fitting the vla surface brightness and the tilt of the @xcite [ o iii ] profile , we get a good fit with an ellipsoid of axial ratio 1.56 , with the near end tilted up ( eastward ) by 9.7@xmath4 .
this is the same direction of tilt shown in fig . 4 of @xcite .
we were guided by the c ii ] profile in setting the central velocity and spatial extent of this fit , since the relatively low resolution of the [ o iii ] profile provided less constraint .
[ bryce_fit ] shows this predicted profile superimposed on the @xcite data .
while no simple ellipsoidal model seems able to fit these complex data , we feel that this is the best compromise , since it ( a ) fits the vla surface brightness , ( b ) produces a front / back brightness ratio in the same sense as that seen in both the c ii ] and [ o iii ] profiles , ( c ) fits the central velocity seen in the c ii ] profile , and ( d ) fits the direction and magnitude of the tilt seen in the [ o iii ] line .
the problem is that this model predicts a tilt in the opposite sense to that seen in the c ii ] line remains . in view of the low s / n of this profile ,
however , its tilt is uncertain .
also , the fact that the ellipsoidal models are only slightly inclined to the line of sight means that irregularities in the shape of the shell can easily change the tilt of the line profiles .
the ultimate goal of our kinematic model is to relate the tangential expansion , as determined by the linear fit in fig .
[ slitfit ] , to the doppler splitting seen in the echelle spectrum . the model discussed above
a tilted ellipsoid with a 1.56 axial ratio has a tangential to radial velocity ratio of 0.667 .
this reduces our estimate of the distance to bd+30 to 1.20 kpc .
the error attached to this distance is hard to estimate . the formal error in the linear fit to the astrometric shifts along the slit is less than 10% , and the error in the expansion velocity smaller yet .
this implies an error of about 10% in the distance , but _ only if we have an accurate kinematic model .
_ it is clear that the shape and kinematics have an important but somewhat uncertain impact on our derived distance , almost surely reducing the value significantly below the spherical value , but by an amount that must remain somewhat uncertain .
progress might be made by a program of extensive kinematic mapping with multiple narrow slits preferably in the infrared to minimize the effects of extinction by internal dust .
we should also bear in mind that all the foregoing analysis is based on the assumption that the expansion of the nebula is radial , and with a velocity proportional to the distance from the center such an expansion will preserve the shape of the nebula . while this is a reasonable and conservative assumption , it is not hard to imagine that there may be more material around the waist of the nebula ( the plane perpendicular to the major axis ) which may impede expansion in that direction .
such a situation would reduce the derived distance to the nebula still further .
we have used pairs of hst wfpc images separated by 5.663 years to produce detailed maps of the angular expansion of bd+30 .
images in both h@xmath0 and [ n ii ] show similar patterns of expansion . near the northeastern corner ,
the direction deviations of expansion flow are very regular , and the magnitudes of expansion are also close to each other . if we look at the opening near the northeastern corner , the movements of the material in the inner part tends to move toward the opening , not always along the radial direction .
because measurements were performed on small independent regions , this can not be a systematic error .
it appears that the material in the inner part is trying to go out through the opening , which has relatively low density , instead of moving along a radial direction .
this probably shows that the inner material has a higher velocity than the outer shell .
the west parts of the image do not have very regular patterns over a large scale , but we can still notice some systematic movements .
for example , the expansions along the southwestern direction going outward are almost all along one direction with same magnitude , and there is another convergence center at the southwestern part of the image .
the velocities are particularly large at the outer edge of the southwest quadrant .
it is interesting that the largest proper motions are approximately along the axis of the bipolar outflow as delineated by co `` bullets '' imaged by @xcite , rather than along the ( ew ) major axis .
apparently , measurements of the angular expansion can reveal the `` true '' bipolar axis , consistent with ground - based kinematics .
overall , our measurements are more secure than earlier results , since the high resolution of these images allows us to measure the displacements of small , discrete features in contrast to the radio work , which could only measure the global expansion . a comparison of proper motions listed in table 2 with the radio expansion measured by @xcite shows a good agreement , but both measurements are significantly larger than the radio expansion determined by @xcite .
the difference between these two radio proper motion measurements arises from a correction used by @xcite to account for the decreasing surface brightness of bd+30 as it expands at an assumed constant luminosity .
the present measurements confirm the validity of this correction , which must be applied if an optically thick nebula is observed with an instrument whose spatial resolution is comparable to the size of the object .
while the hst spatial resolution is superior with respect to the existing radio observations , a rather remarkable agreement with measurements of @xcite demonstrate that ground based radio observations can provide reliable proper motions for compact pne .
although our stis spectra have rather poor s / n , their high spatial resolution gives us a precise measure of the expansion of the brightest part of the nebular shell in a well defined location .
this velocity data can be combined with the astrometric shifts along the same cut to derive a distance . here ,
the greatest uncertainty is the shape and kinematics of the nebula : how elongated is the shell and how uniform and radial is the expansion ?
we have used the radio surface brightness to address the shape , but we have had to assume uniform radial expansion even though our maps of angular expansion show clear non - radial motions is some sectors .
still , we feel that our resulting distance determination of 1.2 kpc is better justified than previous distance estimates , and offers an improvement over the most recent value of @xmath21 kpc found by @xcite from radio expansion measurements combined with the ground - based kinematic data . at a distance of 1.2 kpc ,
the luminosity of the central star of bd+30 is equal to 4250 l@xmath5 , following the recent determination of @xcite based on modeling of the stellar spectrum .
( this is 60% lower than the luminosity based on the earlier model of @xcite . )
most of the stellar radiation is absorbed within the nebula , primarily by dust in the neutral halo ( the infrared luminosity is 2000 l@xmath5 ) and in the photoionized shell ( nebular models indicate that @xmath22 l@xmath5 is required to account for the observed radio continuum ) .
the mostly neutral nebular mass is equal to 0.13 m@xmath5 , based on the [ c ii ] 158 @xmath23 m line luminosity derived from the infrared space observatory ( iso ) measurements of this line flux @xcite and the 1.2 kpc distance . in terms of its mass and luminosity , bd+30 is not exceptional among galactic pne .
but its central wolf - rayet wc star is hydrogen poor and iso observations revealed the presence of crystalline silicate dust in this carbon - rich nebula @xcite , signaling a recent change from the oxygen- to carbon - dominated chemistry .
the origin of these rapid abundance variations is not understood at present .
one promising hypothesis involves a final thermal pulse @xcite which occurred at most a few thousand yr ago , shortly before bd+30 left the asymptotic giant branch phase of evolution . with the fairly well determined distance , perhaps this bright and well - observed pn
can provide a stringent test of various hypotheses invoked to explain the origin of wc stars and the presence of abundance inhomogeneities in pne .
support for this work was provided by nasa through grants go-08116.01 - 97a and go-08116.02 - 97a from the space telescope science institute , which is operated by the association of universities , inc .
, under nasa contract nas5 - 26555 .
we would like to thank dr .
myfanwy bryce for supplying the data from @xcite used in fig .
[ bryce_fit ] .
to the extent that the temperature in the nebular gas is constant , and if we neglect the secondary effects of diffuse radiation produced within the gas , the emission in a hydrogen recombination line or in the free - free continuum will depend only upon the flux of stellar ionizing radiation that is absorbed in the volume of gas . consider a nebular shell which is optically thick to ionizing radiation but which is geometrically thin .
then , the emission from the surface of the shell will simply be proportional to the solid angle of the surface element as seen from the central star .
consider an element of the shell @xmath24 which is inclined so that the angle between the normal to the surface and our line - of - sight is @xmath25 .
let @xmath0 be the angle between the radial vector from the star and the normal to @xmath24 .
then @xmath24 will appear to have a surface brightness @xmath26 of @xmath27}^2\ ] ] where @xmath28 is the total flux from the whole nebula , @xmath29 is the distance of @xmath24 from the central star , and @xmath30 is the distance from the nebula to the observer . near the central star , where @xmath31 , the ratio of @xmath26 to @xmath28 provides the distance of the surface from the star ( in angular units : ( r / d ) radians ) .
one complication is that we see both the front and back sides and must separate these contributions or assume symmetry . from the central surface brightness of the radio frequency images ,
it is apparent that the nebula is elongated along the line - of - sight , since the central intensity is substantially less that would be the case for a sphere .
we find ( r / d ) to be at least @xmath32 ; extrapolation to zero radius of the flux in rings about the center yields @xmath33 .
consider a spherical coordinate system whose origin is the central star , with the z - axis directed towards the observer .
call the azimuthal angle @xmath34 and the angle between the z - axis and the vector to a point on the surface @xmath35 .
the flux in any pie - shaped section of the image , bounded by angles @xmath36 and @xmath37 , must be @xmath38 , because that part of the image arises from the absorption of the fraction of the stellar flux emitted into the volume bounded by the planes @xmath39 and @xmath40 .
in fact , this provides a check on our assumption that the nebula is optically thick and that diffuse nebular radiation is not important : equal pie - shaped segments of the radio image should have the same flux .
this check is satisfied to within 5% , except for a sector at p.a .
@xmath41 , where there is deficit of 11% .
( we use the radio images because the h@xmath0 images are affected by dust absorption / scattering . )
unfortunately , the angular expansion vanishes near the center of the nebula ; what we want is information on the shape of nebula where we have measured the angular expansion .
let @xmath42 be the ( projected ) angular distance of a bit of the surface @xmath24 from the star .
then the flux in the image segment between @xmath43 $ ] and @xmath44 $ ] is equal to the product @xmath45 there are two contributions to @xmath26 : radiation from the front and back surfaces .
we assume that we can separate them or that they are equal .
so @xmath26 here refers just to the front or back surface . the emitting piece of surface is located at the ( unknown ) angle @xmath35 from the star .
the flux emitted by the surface is @xmath46 , where @xmath47 is the solid angle subtended by the surface as seen from the star .
@xmath48 , so that the flux is @xmath49\sin(\theta)~d\phi~d\theta$ ] . by equating fluxes
we can write @xmath50 consider a thin segment of the nebula between @xmath34 and @xmath51 . integrate the expression above from an angle @xmath35 , corresponding to some impact parameter @xmath42 , to @xmath52 , corresponding to @xmath53 .
( @xmath52 is the plane passing through the star and perpendicular to the line - of - sight . if the nebula is expanding radially , then this is the plane with zero radial velocity . )
we find that @xmath54 the expression on the right hand side is just the integrated flux from the azimuthal segment ( from back or front shell only ) , divided by @xmath55 .
we use this relation to find @xmath56 . now
@xmath57 and @xmath58 , so @xmath59 . the p - z plot is the cross - section of the nebula in the @xmath34-plane . as a simple example , let the nebula be an ellipsoid of revolution about the z axis with the minor axis @xmath60 in the plane of the sky and the major axis ( z - axis ) @xmath61 along the line of sight .
then the radius vector from the star to a point on the surface is @xmath62 in this expression , @xmath26 is negative for values of @xmath35 corresponding to the back side of the shell . even for modest @xmath67 and @xmath69 , the effects of @xmath25 and @xmath29 may combine to produce large front - to - back shell brightness ratios . ccccc h@xmath0 f656n & 200,300 & 115 40 58.8 & 1994 - 03 - 06 & 5403 + h@xmath0 f656n & 3 @xmath9 160 & 249 04 05.9 & 1999 - 11 - 04 & 8116 +
[ n ii ] f658n & 200,300 & 115 40 58.8 & 1994 - 03 - 06 & 5403 + [ n ii ] f658n & 3 @xmath9 160 & 249 04 05.9 & 1999 - 11 - 04 & 8116 + lcccc radius & 2.32 & 1.84 & 2.32 & 1.84 + magnification ( mas yr@xmath3 ) & @xmath70 & @xmath71 & @xmath72 & @xmath73 + line magnification ( mas yr@xmath3 ) & @xmath74 & @xmath75 & @xmath76 & @xmath77 + feature shifts ( mas yr@xmath3 ) & @xmath78 & @xmath79 & @xmath80 & @xmath81 + | the wfpc2 camera aboard the _ hubble space telescope _ was used to obtain images of the planetary nebula at two epochs separated by 5.663 years .
the expansion of the nebula in the h@xmath0 and [ ] bands has been measured using several methods .
detailed expansion maps for both emission lines were constructed from nearly 200 almost independent features .
there is good agreement between the ( independent ) h@xmath0 and [ n ii ] proper motions .
there are clear deviations from uniform radial expansion , with higher expansion rates in regions where the shell is faintest , such as the south - west quadrant .
the space telescope imaging spectrograph ( stis ) was used to obtain echelle spectra in the c ii ] @xmath12326 multiplet and the [ o ii ] @xmath12470 doublet , providing well - resolved expansion velocities at two position angles . from the c ii ] lines , we find that the central velocity split is @xmath236.3 km s@xmath3 at a position angle of 99@xmath4 , and @xmath233.5 km s@xmath3 at p.a .
25@xmath4 .
the fainter [ o ii ] doublet does not appear to differ from the c ii ] multiplet . to determine the distance of bd+30@xmath43639 by comparison of the angular expansion and the spectroscopically determined radial expansion
, we must address the problem of the three dimensional shape of the nebula .
we measured the angular expansion along the position of the 99@xmath4 echelle slit , finding displacements of 4.25 mas yr@xmath3 at the shell edge ( 247 from the center ) .
if the nebula were spherical , this would imply a distance of 1.80 kpc .
but there is evidence that the nebula is elongated along the line of sight , which suggests that the actual distance is less .
radio continuum images from 5 and 15 ghz vla observations provide information on the extent of the radial elongation .
we fit the radio brightness variation and the echelle data by approximating the nebula as an ellipsoid , also making use of the ground - based echelle spectra reported by bryce & mellema ( 1999 , mnras , 309 , 731 ) .
our model has an axial ratio of 1.56 , is inclined to the line of sight by 97 , and exhibits an expansion in the plane of the sky which is 2/3 that in the radial direction , leading to a distance of .
not all the kinematic data fits this simple model , so the distance must still be regarded as uncertain .
based on the recent model atmosphere of @xcite , a distance of 1.2 kpc implies a stellar luminosity of 4250 l@xmath5 .
the kinematic age of the nebula , @xmath6 , varies somewhat from region to region .
a good average value is 800 years , while the expansion along the position of the 99@xmath4 echelle slit gives about 600 years . |
production and study of the properties of isotopically enriched substances are one of the crucial problems of contemporary material science involving in itself the increasing number of researchers . owing to the achieved level of purity of many substances , their properties
are substantially determined by the isotopic composition .
development of the technology of high - purity monoisotopic single crystals production allows one to study the influence of the isotopic composition on physical properties , in particular on the electronic , phonon , and raman spectra and the refractive index dispersion .
this is important both for verification of existing theoretical models and filling them with more correct values of used parameters , and for calculation of parameters of next - generation devices developed on their basis .
crystalline silicon is known to be the basic material of microelectronics and solar energetics . besides , it is also an attractive optical material with high transmission in the near- to mid - ir spectral range . in recent years
silicon photonics grows rapidly , which is expected to find application in future optical fiber communication and information transfer systems .
monoisotopic silicon is considered to be the basic component of quantum computers @xcite . formerly
silicon refractive index was measured repeatedly in various spectrum ranges by many laboratories and companies ( see e.g. @xcite ) .
the papers @xcite seem to be the first in this field .
the analysis of more than 50 papers made in 1980 @xcite showed the data available to be scattered widely , but allowed to derive certain averaged spectral dependence for silicon refractive index which is rather often referred to .
unfortunately , in the majority of published papers devoted to silicon refractive index measuring , the crystals used for this purpose are characterized neither in purity , nor in conductivity , both influencing the measurement results . in this paper
we present the results of measuring the room - temperature transmission spectra in the short - wavelength edge and phonon absorption regions , the raman spectra excited by the 514.5 nm ar laser line and the refractive index spectral dependence in the , , and far - ir spectral ranges in silicon single crystals with natural isotopic composition , @xmath3 , and silicon isotopes @xmath0 , @xmath1 , @xmath2 with the basic isotope content above 99.9 at.% prepared by the same technology .
a short note on fabrication of these single crystals and refractive index measurements in the @xmath8 m range has been published recently @xcite .
polycrystalline silicon was prepared by thermal decomposition of monosilane enriched in one of the silicon isotopes as described in @xcite .
the decomposition was proceed on a graphite substrate surface at 800 @xmath9c .
after a few mm thick layer was deposited , polycrystalline silicon was separated from the graphite substrate , remelted to form a rod , and purified by 10 float zoning passes . then this monoisotopic silicon rod was used as a substrate to deposit the polycrystalline silicon layer till an ingot of a required diameter was obtained . to minimize the effect of isotopic dilution by a seeding material , a special growth procedure @xcite was used .
preliminarily a single - crystal seed with the isotopic composition gradient in length was grown by drawing at the end of mono - isotopic silicon ingot of a cylindrical site with a diameter not exceeding that of a seed from silicon of natural isotopic composition .
the composition of grown single crystal seed was almost identical to the isotopic composition of polycrystalline ingot which the single crystal was grown from .
single crystals were grown in the @xmath10 $ ] direction by float zoning in a high - purity argon atmosphere .
the same procedure was used to prepare a single crystal of natural isotopic composition , @xmath3 : @xmath0 ( 92.23% ) , @xmath1 ( 4.68% ) , and @xmath2 ( 3.09% ) @xcite .
c||d .. 5|d .. 5|d .. 5 & + single & + & & & + crystal & & & + & & & + @xmath0 & 99.9934 & 0.00637 & 0.00023 + @xmath1 & 0.026 & 99.919 & 0.055 + @xmath2 & 0.005 & 0.021 & 99.974 + [ tab:1 ] the isotopic composition of the crystals was determined by mass spectrometry @xcite with an accuracy better than 0.01 at.% ( table [ tab:1 ] ) .
comparison with the data reported in @xcite indicates that our crystals are more isotopically enriched and offer higher chemical purity and resistivity .
structural perfection of the crystals was evaluated using x - ray diffraction and selective etching .
no orientation disorder , low - angle boundaries or twinning regions were detected in the single crystals grown .
the half - width of their rocking curves was @xmath11 .
the room - temperature resistivity of the crystals was found to be 100200 @xmath12 . according to mass spectrometric measurements
, the contents of 72 impurities was below the mass spectrometry detection limit ( @xmath13 at.% ) @xcite .
@xmath14 mm - thick plates were cut perpendicularly to the growth axis of the cylindrical samples prepared . then the plates were ground and polished with various flatness degrees for measuring the transmission spectra without interference at 2 @xmath4 resolution ( to analyze the presence of impurity absorption bands and intrinsic bands shift in the two - phonon absorption range ) and with interference at 0.1 @xmath4 resolution ( to measure the refractive index spectral dependence ) .
samples for the refractive index measurements had the surface flatness better than @xmath15 of newton s fringe ( @xmath16 nm ) , parallelism of sides better than 1 , and the surface finish class of 60/40 .
the transmission spectra of crystalline silicon without interference were measured on lambda 900 spectrophotometer in a short - wavelength range ( @xmath17 m ) and on bruker ifs-113v fourier - transform vacuum spectrometer in a longer wavelength range ( @xmath18 m ) .
the analysis on the presence of impurity absorption bands in the transmission spectra measured in conformity with the astm standards f-1188 and f-1391 @xcite proved oxygen and carbon contents to be less than @xmath19 @xmath20 in all the samples under investigation . in the
1 to 3 mm thickness crystals the tendency to the transmission edge shift to higher frequencies with isotope atomic mass increasing was observed in the short - wavelength transmission edge region at room temperature .
however , the shift value of @xmath21 was comparable with the experimental error .
the @xmath22 dependence of the band gap , @xmath23 , was shown in @xcite to be more probable for @xmath24 k , since due to lattice expansion and electron - phonon interaction @xmath23 decreases linearly with @xmath25 growing in the room - temperature region ( @xmath26 k ) .
measurement at 16 k of the transmission spectra of our silicon samples showed that with atomic mass increasing the absorption edge shifts to the high - energy region .
the shift was found to be @xmath27 mev in @xmath0 and @xmath2 relative to @xmath1 . the indirect transition energy determined from the optical absorption edge
was found to depend on the average atomic mass as @xmath22 .
the shift @xmath28 mev measured in the present work at 16 k turned out to be somewhat higher compared with 1 mev derived from the luminescence spectra in @xcite . in the phonon absorption region ( @xmath29
m ) a shift of all absorption bands to longer wavelengths was observed , which had been seen previously in isotopically enriched silicon single - crystal samples with a smaller isotopic enrichment @xcite .
the raman spectra of all produced single crystals were measured on a t64000 triple spectrograph in a backscattering geometry at room temperature .
the 514.5-nm argon laser line was used as an incident radiation .
the scattered radiation was detected by a liquid nitrogen - cooled ccd matrix .
the spectra were gathered in 0.7 @xmath4 steps .
spectrograph was calibrated using radiation in ar laser and hg lamp lines .
the raman band corresponding to first - order scattering by @xmath30 symmetry phonons was observed in spectra of all the single crystals ( showed as circles in fig .
[ fig:1a ] ) .
the band turned out to be well fitted by a lorentzian with width @xmath31 @xmath4 for @xmath0 , @xmath1 , @xmath2 and 3.7 @xmath4 for @xmath3 ( solid lines in fig .
[ fig:1a ] ) .
this is an evidence of high structural perfection of the mono - isotopic single - crystal samples .
the peak position , @xmath32 , of this raman band is plotted vs. atomic mass , @xmath33 , in fig .
[ fig:1b ] for all silicon single crystals .
the dashed line represents the best fit of the experimental phonon frequency values for all the single - crystal samples by the @xmath34 @xmath4 dependence valid in harmonic approximation . the root - mean - square deviation , @xmath35 , of experimental @xmath32 points from the fitted line turned out to be small for the mono - isotopic samples .
again this is an indirect evidence of high structural perfection of those .
the raman band peak position for @xmath3 sample falls slightly above the fitted line , the band being wider than in mono - isotopic samples .
this is attributable to both anharmonicity related to different atomic masses of isotopes and to worse structural perfection of @xmath3 crystal in comparison with the mono - isotopic ones ( see further ) .
to measure refractive index we used the improved method of interference refractometry @xcite . in this method
firstly the order , @xmath36 , of the interference maximum is determined with the absolute accuracy from the peak position , @xmath37 , in the interference transmission spectra of plane - parallel plates cut from the studied material , and then the refractive index value , @xmath38 , is calculated all over the interference transmission range measured using the expression @xmath39 the measurements were performed for plane - parallel plates of three thicknesses ( from 0.8 to 1.2 mm ) cut from silicon samples of all the isotope compositions using bruker ifs-113v fourier - transform vacuum spectrometer in @xmath40 @xmath4 wavenumber range ( @xmath41 m wavelength range ) with 0.1 @xmath4 resolution .
the diameter of a radiation beam transmitted through the samples did not exceed 3 mm .
at least twenty data points were obtained for each interference maximum .
the plates thickness was measured on holographic submicrometer with the accuracy better than 0.05 m. during transmission spectra and thickness measurements sample temperature was hold to be @xmath42 @xmath9c . ,
( b ) @xmath3 , ( c ) @xmath1 , ( d ) @xmath2 ; the literature data for silicon crystals with natural isotopic composition ( i ) @xcite , ( h ) @xcite , ( g ) @xcite , ( e ) @xcite , ( f ) @xcite . ]
[ fig:2 ] shows the refractive index spectral dependences over the entire measurement range for all the single crystal samples studied .
as well the most frequently cited literature data for silicon single crystals with a natural isotopic composition @xcite are shown . as evident from fig .
[ fig:2 ] , the scatter of the refractive index data for silicon with a natural isotopic composition exceeds considerably the difference in the refractive index values of mono - isotopic silicon single crystals . with silicon isotope atomic mass increasing , the refractive index spectral dependences measured by us far from electron and phonon transition edges shifts downwards .
as an example , the refractive index values of the silicon single crystals are listed in table [ tab:2 ] for 1.5 , 2 and 5 m wavelengths .
c||d .. 5|d .. 5|d .. 5 & & & + single crystal & & & + & & & + @xmath0 & 3.48207 & 3.45260 & 3.42194 + @xmath3 & 3.48191 & 3.45243 & 3.42177 + @xmath1 & 3.48171 & 3.45224 & 3.42161 + @xmath2 & 3.48112 & 3.45168 & 3.42107 + [ tab:2 ] it is known that two - phonon lattice absorption bands at 16.4 m ( 610 @xmath4 ) with the maximum intensity of 9 @xmath4 are observable in the @xmath43 m ( @xmath44 @xmath4 ) spectral range in the spectra of @xmath45 mm - thick silicon samples . according to @xcite , those should give rise to well - distinguished features in the refractive index dispersion .
the transmission spectra measured with 2 @xmath4 resolution for all 1.2 mm - thick silicon crystals are shown in fig .
[ fig:3 ] ( a ) .
[ fig:3 ] ( b ) presents the variation of the crystals refractive index measured by us in this spectral region . the absorption band shift to longer wavelengths with silicon isotope atomic mass increasing and the anomalous behavior of refractive index dispersion well - conforming to the classic lorentz theory
are clearly seen .
as far as we know , we managed to observe such a refractive index spectral behavior in the region of so low absorption ( less than 10 @xmath4 ) for the first time .
for example , in @xcite ( line e in fig .
[ fig:2 ] ) the interference refractometry method was used as well to measure the refractive index of n - type silicon ( phosphorus - doped with @xmath46 @xmath47 resistivity ) in @xmath48 m wavelength range at 26 @xmath9c .
however , no peculiarities in the two - phonon absorption region were revealed . in the first approximation ,
the uncertainty in the refractive index values was attributed by the authors of @xcite to the uncertainty in sample thickness ( 1 part per @xmath49 ) , which , in our opinion , is quite enough to observe the changes in the spectral dependence of silicon refractive index shown in fig.[fig:3 ] ( b ) . , @xmath1 , @xmath2 , and @xmath3 samples in the ir region of phonon absorption peak .
+ ( b ) refractive index spectral dependence of the same single crystals . ] the refractive index values calculated from the transmission spectra are conventionally approximated by analytical formulas allowing one to preserve the experimental accuracy of refractive index determination .
the sellmeier or herzberger formulas are often used to approximate the refractive index spectral dependence .
however , our studies show that the application of generalized cauchy dispersion function @xcite allows us to reduce the root - mean - square deviation of the approximation from the calculated refractive index values by several times .
therefore we used it to fit the experimental data : @xmath50 here @xmath51 is the wavenumber , @xmath51 [ @xmath4]@xmath53 [ m ] .
c||d .. 12|d .. 12|d .. 12|d .. 12 & + parameter & & & & + & & & & + & & & & + @xmath54 & 4.03488 10 ^ -27 & 0.0 & 0.0 & -4.47129 10 ^ -20 + @xmath55 & 1.05838 10 ^ -18 & 0.0 & 8.07309 10 ^ -11 & 3.89667 10 ^
-14 + @xmath56 & 1.42643 10 ^ -9 & -1.75666 10 ^ -6 & -1.18530 10 ^ -4 & -1.00590 10 ^ -8 + @xmath57 & 3.41621 & 6.74524 & 68.44582 & 3.41782 + @xmath58 & -34.48103 & -2.35655 10 ^ 6 & -1.58031 10 ^ 7 & -44.46299 + @xmath59 & 0.0 & 7.39436 10 ^ 11 & 1.43557 10 ^ 12 & 0.0 + @xmath60 & 0.0 & -8.68186 10 ^ 16 & 0.0 & 0.0 + & & & & + @xmath61 & 2.0 10 ^ -5 & 3.0 10 ^ -5 & 9.0 10 ^ -5 & 1.1 10 ^ -4 + [ tab:3 ] c||d .. 12|d .. 12|d .. 12|d .. 12 & + parameter & & & & + & & & & + & & & & + @xmath54 & 4.16405 10 ^ -27 & 0.0 & 0.0 & 0.0 + @xmath55 & 1.03238 10 ^ -18 & 0.0 & 1.45163 10 ^ -10 & 4.61074 10 ^ -15 + @xmath56 & 1.42694 10 ^ -9 & -1.35827 10 ^ -6 & -2.03966 10 ^ -4 & -5.60442 10 ^ -10 + @xmath57 & 3.41588 & 5.89678 & 110.64230 & 3.41637 + @xmath58 & -29.06141 & -1.69322 10 ^ 6 & -2.49974 10 ^ 7 & 7.04348 + @xmath59 & 0.0 & 5.12630 10 ^ 11 & 2.18071 10 ^ 12 & -7.25535 10 ^ 5 + @xmath60 & 0.0 & -5.81255 10 ^ 16 & 0.0 & 0.0 + & & & & + @xmath61 & 2.0 10 ^ -5 & 3.0 10 ^ -5 & 1.0 10 ^ -4 & 1.0 10 ^ -4 + [ tab:4 ] c||d .. 12|d .. 12|d .. 12|d .. 12 & + parameter & & & & + & & & & + & & & & + @xmath62 & 0.0 & 1.58348 10 ^ -28 & 0.0 & 0.0 + @xmath63 & 0.0 & -3.70584 10 ^ -22 & 0.0 & 0.0 + @xmath54 & 3.87178 10 ^ -27 & 3.45410 10 ^ -16 & 0.0 & -5.25168 10 ^ -21 + @xmath55 & 1.06782 10 ^ -18 & -1.60314 10 ^ -10 & 1.37221 10 ^ -10 & 1.67675 10 ^ -14 + @xmath56 & 1.42449 10 ^ -9 & 3.70631 10 ^ -5 & -1.87828 10 ^ -4 & -6.42102 10 ^ -9 + @xmath57 & 3.41535 & 0.0 & 99.59448 & 3.41676 + @xmath58 & -50.69159 & 0.0 & -2.18377 10 ^ 7 & -43.76842 + @xmath59 & 0.0 & 0.0 & 1.85529 10 ^ 12 & 0.0 + & & & & + @xmath61 & 2.0 10 ^ -5 & 3.0 10 ^ -5 & 5.0 10 ^ -5 & 1.0 10 ^ -4 + [ tab:5 ] c||d .. 12|d .. 12|d .. 12|d .. 12 & + parameter & & & & + & & & & + & & & & + @xmath54 & 3.85272 10 ^ -27 & 0.0 & 0.0 & 0.0 + @xmath55 & 1.08707 10 ^ -18 & 0.0 & 1.06100 10 ^ -10 & 1.10503 10 ^ -14 + @xmath56 & 1.42510 10 ^ -9 & -1.38069 10 ^ -6 & -1.54334 10 ^ -4 & -4.28163 10 ^ -9 + @xmath57 & 3.41618 & 6.01658 & 87.37999 & 3.41725 + @xmath58 & -96.96621 & -1.82986 10 ^ 6 & -2.02501 10 ^ 7 & -51.15422 + @xmath59 & 0.0 & 5.71062 10 ^ 11 & 1.82698 10 ^ 12 & -1.93263 10 ^ 5 + @xmath60 & 0.0 & -6.67343 10 ^ 16 & 0.0 & 0.0 + & & & & + @xmath61 & 2.0 10 ^ -5 & 2.0 10 ^ -5 & 6.0 10 ^ -5 & 1.3 10 ^ -4 + [ tab:6 ] due to the presence of phonon absorption lines , the range of refractive index approximation was subdivided into four parts .
this allowed us to reduce considerably the root - mean - square deviation between the measured and approximated refractive index values and to bring it to the level not exceeding @xmath64 .
the coefficients of cauchy approximation ( [ ( 2 ) ] ) for refractive index of all silicon single crystals corresponding to the best fit to the experimental results , are given in tables [ tab:3][tab:6 ] .
the values of mean - square deviation for all approximation ranges are listed at the foot of the tables . as an example , the spectral dependence of the difference between the calculated refractive index for @xmath2 single crystal and its approximated values is shown in fig .
[ fig:4 ] .
it is seen that the deviation of the approximation from the calculated refractive index values does not exceed @xmath65 in a sufficiently wide range .
however , with the wavelength exceeding 16 m the accuracy of peak position determination decreases owing to noticeable growth of noise in the transmission spectra . accordingly
, the accuracy of the refractive index calculation is reduced . comparing our results to the data published , we found that the refractive index values closest to ours for natural silicon are given in @xcite . in that work measurements
were carried out by the prism method with @xmath64 relative accuracy at several wavelengths in the @xmath66 m range at temperature from 20 to 300 k. the results obtained were approximated by the sellmeier formula . , and approximated , @xmath67 , refractive index for @xmath2 single crystal . ] the observed shift of the dispersion curve is easily understood using classic models of a lattice of oscillators with dipolar coupling @xcite . neglecting the spatial dispersion the dielectric constant of a cubic lattice
is expressed in this model by the equation @xmath68 where @xmath33 is the atomic mass of @xmath69 isotope , @xmath70 is the band gap ( fundamental absorption short - wavelength edge energy ) , @xmath71 is the phonon frequency ( fundamental absorption long - wavelength edge energy ) , @xmath72 and @xmath73 are the model parameters representing two types of oscillators corresponding to the electron and phonon fundamental absorption edges , respectively . with no absorption , the refractive index equals to @xmath74 .
the fundamental absorption edge energies depend on isotopic composition and for an isotopically pure crystal can be expressed as @xcite : @xmath75 \\ \omega\left(m_{i}\right ) & = & \omega_0\left(\frac{m_{nat}}{m_{i}}\right)^{1/2}\end{aligned}\ ] ] where @xmath76 is the average atomic mass of natural silicon , @xmath77 is phonon bose distribution function , @xmath78 is the model parameter . according to @xcite , @xmath79 @xmath4 , @xmath80 @xmath4 .
the oscillator parameters , @xmath72 and @xmath73 , depend on isotopic composition as well .
first , both parameters depend on the lattice constant as @xmath81 and @xmath82 is known to change with isotopic composition ( see e.g. @xcite ) . however , this dependence is negligible in the qualitative model under consideration .
second , in the case of silicon the phonon - related parameter , @xmath73 , describes the ir absorption due to two - phonon transitions and hence depends on the atomic mass as @xmath83 @xcite .
therefore @xmath84 where @xmath85 , @xmath86 .
thus , with si atomic mass increasing , the electronic absorption edge shifts to shorter wavelengths , and the phonon absorption edge shifts to longer wavelengths . as a result
, the dispersion curve of the refractive index in the transmission window should shift downwards without considerable changes in shape , as observed in our experiments . for optimal and differing by @xmath87 values of @xmath88 . ]
to measure the refractive index precisely by interference refractometry , the orders , @xmath36 , of interference maxima should be determined with absolute accuracy . for this purpose
we used the relation following from expression ( [ ( 1 ) ] ) written for two samples of different thickness , @xmath89 and @xmath90 , with interference orders @xmath91 and @xmath92 , respectively : @xmath93 if sample thickness is known with absolute accuracy , the constant in the right side of this equation turns to zero for a certain relation between @xmath91 and @xmath92 .
however , usually the sample thickness can be measured with some finite accuracy resulting in non - zero value of this constant @xcite . moreover ,
if the transmission spectra of the samples of different thickness are measured at different temperatures , spectral dependence of the left side of equation ( [ ( 3 ) ] ) will be affected by the temperature dependence of the refractive index of a material . however ,
if one makes a wrong choice of the maxima orders at least by 1 , the spectral dependence of the left side of equation ( [ ( 3 ) ] ) will explicitly differ from constant as evident from fig .
[ fig:5 ] for two @xmath3 samples .
minimization of the right side of dependence ( [ ( 3 ) ] ) allows us to determine uniquely the maxima orders for each pair of samples . additionally , the procedure of interference orders determination becomes simpler if the formula ( [ ( 1 ) ] ) is considered to be valid not only for the maxima but as well for all other points of the spectral transmission curve for which @xmath94 values are not integer .
one should use a linear approximation for @xmath88 frequency dependence in the interval between two nearest maxima ( according to our calculations such an approximation is valid with the accuracy about @xmath95 for @xmath96 @xmath4 distance between the maxima ) and increase the corresponding order by 1 passing over the next maximum . in this case
the application of the ( [ ( 1 ) ] ) criterion becomes a simple search of orders corresponding to a given frequency in the transmission spectra of sample pairs , since the set of frequency values is the same for all spectra . in the method of interference refractometry
the error of the refractive index calculation depends on determination errors of all the variables occurring in the formula ( [ ( 1 ) ] ) .
if fourier - transform spectrometer with the wavenumber precision better than 0.01 @xmath4 is used , this error is mostly contributed by the sample thickness measurement error , @xmath97 . the corresponding error in silicon
refractive index is equal to @xmath98 .
if @xmath99 m and values of @xmath36 are determined precisely for all interference maxima , one gets @xmath100 on average for the only @xmath101 mm - thick sample used .
using of two and more samples allows us to reduce this error due to the correlation of the refractive index values determined for all samples over the entire measurement range . assuming the refractive index values to be identical for all the samples within the entire measurement range
, we can analyze the influence of accuracy of the used thickness values on the final result of measurements and correct these thickness values .
analysis of the discrepancy between the corrected and initially used thicknesses provides an idea of how well the samples are prepared and measurements are made .
correction of thickness values ( @xmath102 , @xmath103 in the case of three samples ) is performed by orthogonal projecting in the @xmath104 space the point with coordinates equal to these thickness values on a straight line representing solution of the following set of equations for all pairs of samples @xmath105 the refractive index values scatter , @xmath106 , for sample of thickness @xmath102 is obtained immediately from the relation @xmath107 where @xmath108 is the averaged refractive index value , @xmath109 is the difference between the corrected and initial thickness .
table [ tab:7 ] lists sample thickness measured on submicrometer at 21.2 @xmath9c and the thickness values corrected using the interference spectra in the middle ( mir , from 2 to 25 m ) and near ( nir , from 1.06 to 2 m ) infrared ranges , as well as the corresponding difference between the calculated and measured thicknesses .
as is seen from table [ tab:7 ] the thickness correction value turns out to be extremely small for most of the samples not exceeding the error of direct measurement of the thickness .
we estimate the average error of the refractive index determination resulting from the error of the samples thickness determination to be @xmath110 . the contribution of the remaining factors influencing the samples transmission spectra measurements and , accordingly , our calculation of the refractive index performed similarly to paper @xcite , is found to be 10 times less and
therefore is negligible .
averaged between refractive index for @xmath0 samples of different thicknesses . ] unlike the errors of thickness measurement and interference maxima orders determination , the error of the maxima positions determination is the only one exhibiting distinct frequency dependence : the latter error grows with the frequency decreasing ( wavelength increasing ) or in the absorption band vicinity and near the transparency range edges if any in the transmission spectra . to determine the maxima positions
more exactly we restricted the range of the sample thicknesses used and spectrometer resolution to get not less than 20 experimental points between adjacent maxima in the corresponding transmission spectra . to estimate the influence of error of the maxima positions determination on the refractive index calculation , we determined the averaged difference between the refractive index values calculated for samples with different thickness .
as an example , such a difference is presented in fig .
[ fig:6 ] for @xmath0 samples .
c||d .. 4|d .. 4|d .. 4|d .. 3|d .. 3 & & & & + & & & & & + & & & & + & 815.87 & 815.98 & 815.99 & 0.11 & 0.12 + @xmath0 & 1003.25 & 1003.23 & 1003.23 & -0.02 & -0.02 + & 1179.09 & 1179.03 & 1179.02 & -0.06 & -0.07
+ & & & & + & 815.95 & 815.98 & 815.98 & 0.03 & 0.03 + @xmath1 & 1003.23 & 1003.21 & 1003.23 & -0.02 & 0.0 + & 1179.24 & 1179.23 & 1179.22 & -0.01 & -0.02 + & & & & + & 816.09 & 816.12 & 816.13 & 0.03 & 0.04 + @xmath2 & 1003.34 & 1003.41 & 1003.38 & 0.07 & 0.04 + & 1179.54 & 1179.46 & 1179.48 & -0.08 & -0.06 + & & & & + & 815.87 & 815.88 & 815.88 & 0.01 & 0.01 + @xmath3 & 1003.27 & 1003.28 & 1003.28 & 0.01 & 0.01 + & 1179.15 & 1179.14 & 1179.14 & -0.01 & -0.01 + [ tab:7 ]
in this work we have prepared and characterized high - purity single crystals of silicon stable isotopes , @xmath0 , @xmath1 , and @xmath2 with enrichment above 99.9 at.% , and of silicon with natural isotopic composition , @xmath3 .
for the first time the spectral dependences of the refractive index of all single crystals are measured in wide transparency range from 1.06 to about 80 m including the phonon absorption region .
the results of our study prove the refractive index values of mono - isotopic silicon single crystals to differ considerably .
this may provide a possibility of manufacture of multilayer wave - guiding and quantum - sized structures operating in the ir spectrum range from different silicon isotopes to solve various problems of electronics and fiber and integrated optics .
this work was supported by the russian foundation for basic research ( grants no . and no .
) , by the russian federation president s grants council for support to the leading scientific schools of russia ( grant no . ) and by the presidium of the russian academy of sciences basic research projects novel optical materials and principles of basic research in nanotechnologies and nanomaterials
. 99 p.t.greenland , s.a.lynch , a.f.g.vandermeer , b.n.murdin , c.r.pidgeon , b.redlich , n.q.vinh , and g.aeppli , nature 465 , 1057 ( 2010 ) v.g.plotnichenko , v.o.nazaryants , e.b.kryukova , v.v.koltashev , v.o.sokolov , a.v.gusev , v.a.gavva , m.f.churbanov , and e.m.dianov , quantum electron .
40 , 753 ( 2010 ) .
g.g.devyatykh , e.m.dianov , a.d.bulanov , o.yu.troshin , v.v.balabanov , and d.a.pryakhin , doklady chemistry 391 , 204 ( 2003 ) m.f.churbanov , a.d.bulanov , a.p.kotkov , a.m.potapov , o.yu.troshin , a.yu.loshkov , n.d.grishnova , and s.a.adamchik , doklady chemistry 432 , 126 ( 2010 ) a.v.gusev and v.a.gavva , russian federation patent no.2370576 from 20.10.2009 j.r.rosman and p.d.taylor , pure appl .
70 , 217 ( 1998 ) i.d.kovalev , a.m.potapov , and a.d.bulanov , mass spectrometry 1 , 37 ( 2004 ) k.m.itoh , j.kato , m.m.uemura , a.k.kaliteevskii , o.n.godisov , g.g.devyatych , a.d.bulanov , a.v.gusev , i.d.kovalev , p.g.sennikov , h.j.pohl , n.v.abrosimov , and h.riemann , jpn . j. appl .
42 , 6248 ( 2003 ) j.w.ageriii , j.w.beeman , w.l.hansen , e.e.haller , i.d.sharp , c.liao , a.yang , m.w.l.thewalt , and h.riemann , j.electrochem .
152 , g448 ( 2005 ) i.d.kovalev , k.n.malyshev , a.m.potapov , and a.i.suchkov , j. analytical chem . 56 , 437 ( 2001 ) annual book of astm standards , vol.10.05 electronics(ii ) , 1996 m.cardona and m.l.w.thewalt , rev .
modern phys .
77 , 1173 ( 2005 ) p.g.sennikov , t.v.kotereva , a.g.kurganov , b.a.andreev , h.niemann , d.schiel , v.v.emtsev , and h .- j.pohl , semiconductors 39 , 320 ( 2005 ) a.v.vasilev , v.v.voitsekhovskii , and v.g.plotnichenko , high - purity subst . 3 , 39 ( 1991 ) ( in russian ) v.g.plotnichenko , v.o.nazaryants , e.b.kryukova , and e.m.dianov , j. physics d : applied physics 43 , 105402 ( 2010 ) v.o.nazaryants , e.b.kryukova , e.m.gavrishchuk , v.b.ikonnikov , s.m.mazavin , and v.g.plotnichenko , appl .
optics 49 , 4723 ( 2010 ) h.b.briggs , phys .
77 , 286 ( 1950 ) c.d.salzberg and j.j.villa , j. opt .
47 , 244 ( 1957 ) h.h.li , j .phys .
data 9 , 561 ( 1980 ) d.f.edwards and e.ochoa , appl .
optics 19 , 4130 ( 1980 ) b.j.frey , d.b.leviton , and t.j.madison , proc .
spie 6273 , 62732j ( 2006 ) v.g.plotnichenko , v.o.nazaryants , e.b.kryukova , yu.n.pyrkov , e.m.dianov , b.i.galagan , and s.e.sverchkov , inorg . mater .
45 , 322 ( 2009 ) d.y.smith , m.inokuti , and w.karstens , j. phys . : condens .
13 , 3883 ( 2001 ) u.fano , phys .
118 , 451 ( 1960 ) v.m.agranovich and v.l.ginzburg , crystal optics with spatial dispersion and excitons , springer - verlag , berlin , 1984 , 441p .
m.cardona , phys .
( b ) 220 , 5 ( 2000 ) l.f.lastras-martnez , t.ruf , m.konuma , m.cardona , and d.e.aspnes , phys .
b 61 , 12946 ( 2000 ) a.k.ramdas , s.rodriguez , s.tsoi , and e.e.haller , solid state communications 133 , 709 ( 2005 ) c.p.herrero ,
solid state communications 110 , 243 ( 1999 ) k.winer and m.cardona , phys .
b 35 , 8189 ( 1987 ) r.gupta and s.g.kaplan , j. res .
108 , 429 ( 2003 ) | precise measurement of the refractive index of stable silicon isotopes @xmath0 , @xmath1 , @xmath2 single crystals with enrichments above 99.9 at.% and a silicon single crystal @xmath3 of natural isotopic composition is performed with the fourier - transform interference refractometry method from 1.06 to more than 80 m with 0.1 @xmath4 resolution and accuracy of @xmath5 .
the oxygen and carbon concentrations in all crystals are within @xmath6 and the content of metal impurities is @xmath7 at.% .
the peculiar changes of the refractive index in the phonon absorption region of all silicon single crystals are shown .
the coefficients of generalized cauchy dispersion function approximating the experimental refractive index values all over the measuring range are given .
the transmission and raman spectra are also studied . |
secure communication is a topic that is becoming increasingly important thanks to the proliferation of wireless devices . over the years
, several secrecy protocols have been developed and incorporated in several wireless standards ; e.g. , the ieee 802.11 specifications for wi - fi . however ,
as new schemes are being developed , methods to counter the specific techniques also appear . breaking this cycle is critically dependent on the design of protocols that offer provable secrecy guarantees .
the information theoretic secrecy paradigm adopted here , allows for a systematic approach for the design of low complexity and provable secrecy protocols that fully exploit the intrinsic properties of the wireless medium .
most of the recent work on information theoretic secrecy is , arguably , inspired by wyner s wiretap channel @xcite . in this setup
, a passive eavesdropper listens to the communication between two legitimate nodes over a separate communication channel . while attempting to decipher the message , no limit is imposed on the computational resources available to the eavesdropper .
this assumption led to defining * perfect secrecy capacity * as the maximum achievable rate subject to zero mutual information rate between the transmitted message and the signal received by the eavesdropper . in the additive gaussian noise scenario
@xcite , the perfect secrecy capacity turned out to be the difference between the capacities of the legitimate and eavesdropper channels . therefore ,
if the eavesdropper channel has a higher channel gain , information theoretic secure communication is not possible over the main channel .
recent works have shown how to exploit multipath fading to avoid this limitation @xcite .
the basic idea is to opportunistically exploit the instants when the main channel enjoys a higher gain than the eavesdropper channel to exchange secure messages .
this opportunistic secrecy approach was shown to achieve non - zero * ergodic secrecy capacity * even when * on average * the eavesdropper channel has favorable conditions over the legitimate channel .
remarkably , this result still holds even when the channel state information of the eavesdropper channel is not available at the legitimate nodes .
the ergodic result in @xcite applies only to delay tolerant traffic , e.g. , file downloads .
early attempts at characterizing the delay limited secrecy capacity drew the negative conclusion that non - zero delay limited secrecy rates are not achievable , over almost all channel distributions , due to * secrecy outage * events corresponding to the instants when the eavesdropper channel gain is larger than the main one @xcite .
later , it was shown in @xcite that , interestingly , a non - zero delay limited secrecy rate could be achieved by introducing * private key queues * at both the transmitter and the receiver .
these queues are used to store private key bits that are shared * opportunistically * between the legitimate nodes when the main channel is more favorable than the one seen by the eavesdropper .
these key bits are used later to secure the delay sensitive data using the vernam one time pad approach @xcite .
hence , secrecy outages are avoided by simply storing the secrecy generated previously , in the form of key bits , and using them whenever the channel conditions are more advantageous for the eavesdropper . however , this work stopped short of proving sharp capacity results or deriving the corresponding optimal power control policies .
these results can be recovered as special cases of the secrecy outage capacity and power control characterization obtained in the sequel .
in particular , this work investigates the outage secrecy capacity of point - to - point block fading channels .
we first consider the scenario where perfect knowledge about the main and eavesdropper channels are available _ a - priori _ at the transmitter . the outage secrecy capacity and corresponding optimal power control policy is obtained and then the results are generalized to the more practical scenario where only the main channel state information ( csi ) is available at the transmitter .
finally , the impact of the _ private key queue _ overflow on secrecy outage probability is studied .
overall , our results reveals interesting structural insights on the optimal encoding and power control schemes as well as sharp characterizations of the fundamental limits on secure communication of delay sensitive traffic over fading channels .
the rest of this paper is organized as follows.we formally introduce our system model in section [ s : sysmodel ] . in section [ section : capacity
] , we obtain the capacity results for the full and main csi scenarios .
the optimal power control policies , for both cases , are derived in section [ section : power ] .
the effect of key buffer overflow on the outage probability is investigated in section [ s : finitebuffer ] .
we provide simulations to support our main results in section [ section : simulations ] .
finally , section [ s : conclusion ] offers some concluding remarks . to enhance the flow of the paper , the proofs
are collected in the appendices .
we study a point - to - point wireless communication link , in which a transmitter is trying to send information to a legitimate receiver , under the presence of a passive eavesdropper .
we divide time into discrete slots , where blocks are formed by @xmath0 channel uses , and @xmath1 blocks combine to form a super - block .
let the communication period consist of @xmath2 super - blocks .
we use the notation @xmath3 to denote the @xmath4 block in the @xmath5 super - block .
we adopt a block fading channel model , in which the channel is assumed to be constant over a block , and changes randomly from one block to the next . within each block @xmath3 ,
the observed signals at the receiver and at the eavesdropper are : @xmath6 and @xmath7 respectively , where @xmath8 is the transmitted signal , @xmath9 is the received signal by the legitimate receiver , and @xmath10 is the received signal by the eavesdropper .
@xmath11 and @xmath12 are independent noise vectors , whose elements are drawn from standard complex normal distribution .
we assume that the channel gains of the main channel @xmath13 and the eavesdropper channel @xmath14 are i.i.d .
complex random variables .
the power gains of the fading channels are denoted by @xmath15 and @xmath16 .
we sometimes use the vector notation @xmath17 $ ] for simplicity , and also use the notation @xmath18 to denote the set of channel gains @xmath19 observed until block @xmath3 .
we use similar notation for other signals as well , and denote the sample realization sequences with lowercase letters .
we assume that the probability density function of instantaneous channel gains , denoted as @xmath20 , is well defined , and is known by all parties .
we define channel state information ( csi ) as one s knowledge of the instantaneous channel gains .
we define _ full transmitter csi _ as the case in which the transmitter has full causal knowledge of the main and eavesdropper channel gains .
we define _ main transmitter csi _ as the case in which that the transmitter only knows the csi of the legitimate receiver . in both cases , the eavesdropper has complete knowledge of both the main and the eavesdropper channels .
let @xmath21 denote the power allocated at block @xmath3 .
we consider a long term power constraint ( or average power constraint ) such that , @xmath22 for some @xmath23 .
let @xmath24 denote the set of messages to be transmitted with a delay constraint .
@xmath25 becomes available at the transmitter at the beginning of block @xmath3 , and needs to be securely communicated to the legitimate receiver at the end of that particular block .
we consider the problem of constructing @xmath26 codes to communicate message packets @xmath27 of equal size , which consists of : 1 . a stochastic encoder that maps @xmath28 , @xmath29 to @xmath30 based on the available csi , where @xmath31 summarizes the previously transmitted signals , in which case the previous signals are summarized by @xmath32 .
] , and 2 . a decoding function that maps @xmath33 to @xmath34 at the legitimate receiver .
note that we consider the current block @xmath30 to be a function of the past blocks @xmath31 as well .
this kind of generality allows us to store shared randomness to be exploited in the future to increase the achievable secrecy rate .
define the error event with parameter @xmath35 at block @xmath3 as @xmath36 which occurs either when the decoder makes an error , or when the power expended is greater than @xmath37 .
the equivocation rate at the eavesdropper is defined as the entropy rate of the message at block @xmath3 , conditioned on the received signal by the eavesdropper during the transmission period , and available eavesdropper csi , which is equal to @xmath38 .
the secrecy outage event at rate @xmath39 with parameter @xmath35 at block @xmath3 is defined as @xmath40 where the equivocation outage @xmath41 occurs if the equivocation rate at block @xmath3 is less than @xmath42 , and channel outage @xmath43 occurs if channel at block @xmath3 is unsuitable for reliable transmission at rate @xmath39 . defining @xmath44 as the complement of the event @xmath45 ,
we now characterize the notion of @xmath46-achievable secrecy capacity .
rate @xmath39 is achievable securely with at most @xmath46 probability of secrecy outage if , for any fixed @xmath47 , there exist @xmath48 and @xmath0 large enough such that the conditions @xmath49 are satisfied for all @xmath3 , @xmath50 .
we call such @xmath39 an @xmath46-achievable secrecy rate .
note that the security constraints are not imposed on the first super - block .
the @xmath46-achievable secrecy capacity is the supremum of @xmath46-achievable secrecy rates @xmath39 .
the notion of secrecy outage was previous defined and used in @xcite .
however , those works did not consider the technique of storing shared randomness for future use , and in that case , secrecy outage depends only on the instantaneous channel states . in our case ,
secrecy outage depends on previous channel states as well .
note that we do not impose a secrecy outage constraint on the first superblock ( @xmath51 ) .
we refer to the first superblock as an initialization phase used to generate initial common randomness between the legitimate nodes .
note that this phase only needs to appear _ once _ in the communication lifetime of that link .
in other words , when a session ( which consists of @xmath2 superblocks ) between the associated nodes is over , they would have sufficient number of common key bits for the subsequent session , and would not need to initiate the initialization step again .
in this section , we investigate this capacity under two different cases ; full csi and main csi at the transmitter . before giving the capacity results ,
we define the following quantities . for a given power allocation function @xmath21 ,
let @xmath52 and @xmath53 be as follows , @xmath54^+ \label{eqn : rs(t)}\end{aligned}\ ] ] where @xmath55^+= \max(\cdot,0)$ ] .
note that , @xmath56 is the supremum of achievable main channel rates , without the secrecy constraint .
also , @xmath57 is the non - negative difference between main channel and eavesdropper channel s supremum achievable rates .
we show in capacity proofs that the outage capacity achieving power allocation functions lie in the space of stationary power allocation functions that are functions of instantaneous transmitter csi .
hence for * full csi * , we constrain ourselves to the set @xmath58 of stationary power allocation policies that are functions of @xmath59 $ ] . for simplicity , we drop the block index @xmath3 , and use the notation @xmath60 for the stationary power allocation policy . similarly , with * main csi * we consider the power allocation policies that are functions of @xmath61 , and use the notation @xmath62 for the stationary power allocation policy . in both cases , since the secrecy rate @xmath53 , and the main channel rate @xmath52 are completely determined by the stationary power allocation functions @xmath63 and channel gains @xmath64 , we will interchangeably use the notations @xmath65 and @xmath66 .
[ fullcsicapacity ] let the transmitter have full csi . then , for any @xmath46 , @xmath67 , the @xmath46-achievable secrecy capacity is identical to @xmath68}{1-\epsilon } \label{secrecycapacity}\end{aligned}\ ] ] where the set @xmath69 consists of power control policies @xmath60 that satisfies the following conditions .
@xmath70}{1-\epsilon}\right ) & \leq \epsilon
\label{eq : outageconstraint}\\ \expect [ p({{\bf h } } ) ] & \leq p_{\text{avg } } \label{eq : powerconstraint2}\end{aligned}\ ] ] a detailed proof of achievability and converse part is provided in appendix [ a : fullcsi ] . here
, we briefly justify the result . for a given @xmath60 , @xmath71 the supremum of the secret key generation rates within a block that experiences channel gains @xmath64 @xcite .
this implies that the expected achievable secrecy rate @xcite is @xmath72 $ ] without the outage constraint . with the outage constraint , the fluctuations of @xmath73 due to fading are unacceptable , since @xmath73 can go below the desired rate when the channel conditions are unfavorable ( e.g. , when @xmath74 , @xmath75 ) .
hence , we utilize secret key buffers to smoothen out these fluctuations to provide secrecy rate of @xmath72 $ ] at each block .
the generated secrecy is stored in secret key buffers of both the transmitter and receiver , and is utilized to secure data of same size using vernam s one - time pad technique . with the allowable amount of secrecy outages , this rate goes up to @xmath72/(1-\epsilon)$ ] .
the channel outage constraint on the other hand is a necessary condition to satisfy the secrecy outage constraint in due to .
[ e : fullcsi ] consider a four state system , where @xmath76 and @xmath77 takes values from the set @xmath78 and the joint probabilities are as given in table [ ex1:pr ] .
let the average power constraint be @xmath79 , and there is no power control , i.e. , @xmath80 @xmath81
. the achievable instantaneous secrecy rate at each state is given in table [ ex1:rs ] .
according to the pessimistic result in [ 6,8 ] , any non - zero rate can not be achieved with a secrecy outage probability @xmath82 in this case . however , according to theorem 1 , rate @xmath83 can be achieved with @xmath46 secrecy outage probability is continuous , the result similarly applies to discrete @xmath84 as well . ] , since @xmath85=0.8 $ ] .
a sample path is provided for both schemes in figure [ fig : sample_path ] , and it is shown how our scheme avoids secrecy outage in the second block . .@xmath86
[ cols="^,^,^",options="header " , ] here , we find the optimal power control strategy @xmath87 , which solves the optimization problem - .
let us define @xmath88 as the maximum of @xmath89 , and the solution of the following equation @xmath90}{\partial p(h_m)}= \frac{h_m \pr ( h_e \leq h_m)}{1+h_mp(h_m)}-\int_0^{h_m}\left ( \frac{h_e}{1+h_ep(h_m ) } \right)f(h_e)dh_e - \lambda = 0\end{aligned}\ ] ] @xmath88 will replace @xmath91 in the full csi case .
@xmath92 is the unique solution to @xmath93/(1-\epsilon ) \label{moptimalpower:2}\\ & \pr(h_m \geq c ) = 1-\epsilon \label{moptimalpower:3}\\ & \expect[p^{\ast}(h_m ) ] = p_{\text{avg } } \label{moptimalpower:4}\end{aligned}\ ] ] where @xmath94 $ ] is the expected secrecy rate under the power allocation policy @xmath92 .
the proof follows the approach in full csi case , hence we omit the details for brevity . define the sub - problem @xmath95 = & \max_{p(h_m)}\expect \left [ r_s({{\bf
h}},p ) \right ] \label{eq : mainpowercontrol}\\ \textrm{subject to : } & p(h_m ) \geq 0,~ \forall h_m \nonumber\\ & \expect [ p(h_m ) ] \leq p_{\text{avg } } , \label{mainpowerconstraint}\\ & \pr\left(r_m({{\bf h}},p ) < { r}\right ) \leq
\epsilon \label{mainserviceconstraint}\end{aligned}\ ] ] let @xmath96 be the power allocation function that solves this sub - problem .
lemmas [ bmaxlemma ] and [ fullcsiuniquepwr ] also hold in this case .
the only difference is the following lemma , which replaces lemma [ l : fullcsipower ] in full csi .
[ t : maincsipower ] for any @xmath97 , @xmath98 where @xmath99 is a constant that satisfies @xmath100 , and @xmath101 is a constant that satisfies with equality .
the proof is similar to the proof of lemma [ l : fullcsipower ] , and is provided in appendix [ a : maincsipower ] .
the graphical solution in figure [ fig : fullcsipower ] to find @xmath102 also generalizes to the main csi case .
the proofs of the capacity results of section [ section : capacity ] assume availability of _ infinite size _ secret key buffers at the transmitter and receiver , which mitigate the effect of fluctuations in the achievable secret key bit rate due to fading .
finite - sized buffers , on the other hand will lead to a higher secrecy outage probability due to wasted key bits by the key buffer overflows .
we revisit the full csi problem , and we consider this problem at ` packet ' level , where we assume a packet is of fixed size of @xmath0 bits .
we will prove the following result .
[ eq : buffervsoutage ] let @xmath103 .
let @xmath104 be the buffer size ( in terms of packets ) sufficient to achieve rate @xmath102 with at most @xmath105 probability of secrecy outage .
then , @xmath106 + ( c_f^{\epsilon})^2\epsilon(1-\epsilon)}{(\epsilon'-\epsilon)c_f^{\epsilon } } \log\left ( \frac{\var[r_s({{\bf h}},p^{c_f^{\epsilon } } ) ] + ( c_f^{\epsilon})^2\epsilon(1-\epsilon ) } { ( \epsilon'-\epsilon)^2 c_f^{\epsilon } } \right ) } \leq 1 \label{c : asymptoticbuffer}\end{aligned}\ ] ] before providing the proof
, we first interpret this result .
if buffer size is infinite , we can achieve rate @xmath102 with @xmath46 probability of secrecy outage . with finite buffer
, we can achieve the same rate with @xmath105 probability of secrecy outage .
considering this difference to be the price that we have to pay due to the finiteness of the buffer , we can see that the buffer size required scales with @xmath107 , as @xmath108 .
achievability follows from simple modifications to the capacity achieving scheme described in appendix [ a : fullcsi ]
. we will first study the key queue dynamics , then using the heavy traffic limits , we provide an upper bound to the key loss ratio due to buffer overflows .
then , we relate key loss ratio to the secrecy outage probability , and conclude the proof . for the key queue dynamics , we use a single index @xmath109 to denote the time index instead of the double index @xmath3 , where @xmath110 .
we consider transmission at outage secrecy rate of @xmath39 , and use power allocation function @xmath111 , which solves the problem - .
let us define @xmath112 as the key queue process with buffer size @xmath113 , and let @xmath114 .
then , during each block @xmath109 , 1
. the transmitter and receiver agree on secret key bits of size @xmath115 using privacy amplification , and store the key on their secret key buffers .
2 . the transmitter pulls key bits of size @xmath39 from its secret key buffer to secure the message stream of size @xmath39 using one time pad , and transmits over the channel .
as explained in appendix [ a : fullcsi ] .
the last phase is skipped if outage ( @xmath116 ) is declared , which is triggered by one of the following events * channel outage ( @xmath117 ) : the channel can not support reliable transmission at rate @xmath39 , i.e. @xmath118 . *
key outage ( @xmath119 ) : there are not enough key bits in the key queue to secure the message at rate @xmath39 .
this event occurs when @xmath120 . * artificial outage ( @xmath121 ) : outage is artificially declared , even though reliable transmission at rate @xmath39 is possible . due to the definition of @xmath111 , @xmath122 @xmath123 , and the set @xmath124 of events indexed by @xmath109
are i.i.d .
we choose @xmath125 such that @xmath126 is i.i.d . as well , and @xmath127 the dynamics of the key queue can therefore be modeled by @xmath128 note that @xmath129 @xmath123 , due to the definition of @xmath130 .
let @xmath131 be the time average loss ratio over the first @xmath132 blocks , for buffer size @xmath113 , which is defined as the ratio of the amount of loss of key bits due to overflows , and the total amount of input key bits @xmath133 then , we can see that @xmath134 , @xmath135 follows from , , and the fact that @xmath114 .
[ l : stationarity ] @xmath136 converges in distribution to an almost surely finite random variable . the proof is provided in appendix [ app : stationarity ] .
this implies that @xmath137 exists .
now , we provide our asymptotic result for the key loss ratio .
we define the drift and variance of this process as @xmath138 \nonumber\\ & = \expect[r_s({{\bf h}},p^{{r}})]-{r}(1-\epsilon ) \label{mu_rate}\end{aligned}\ ] ] and @xmath139 \nonumber \end{aligned}\ ] ] respectively , where follows from the definition of @xmath140 .
[ t : losspr ] for any @xmath141 , the key loss ratio satisfies the following asymptotic relationship @xmath142 e^{\frac{-2{r}|{\mu}_{{r}}|}{\sigma_{{r}}^2 } } } { { \sigma}_{{r}}^2 } \leq e^{-2m}\end{aligned}\ ] ] the proof is provided in appendix [ app : bufferoverflowproof ] . [ outagelemma ] if @xmath143 , then @xmath105 secrecy outage probability is satisfied .
find @xmath1 such that @xmath144 for any @xmath145 . in 2-index time notation @xmath3 with @xmath146
, it corresponds to @xmath147 , @xmath148 . then .
@xmath149 here , follows from the union bound , and second term follows from the equivocation analysis and in appendix a , which shows that there exists some packet size @xmath0 large enough such that @xmath150 .
equation implies that @xmath105 secrecy outage probability is satisfied .
let @xmath143 .
since @xmath151 and @xmath152 , we have @xmath153 .
this implies that @xmath154 ( since otherwise , key outage probability would be zero ) , which , due to implies @xmath155 & = ( 1- \lim_{t \to \infty}\pr({{{\cal o}}}_{\text{enc}}(t))){r}\nonumber \\ & = ( 1-\epsilon'){r}\label{corr1:ltm}\end{aligned}\ ] ] here , due to the choice of power allocation function @xmath156 , we have @xmath157 = \lim_{t\to\infty}\frac{1}{t}\sum_{t=1}^t r_s(t)$ ] . plugging the result of lemma [ t : losspr ] into , we obtain the required key buffer size to achieve @xmath105 probability of secrecy outage @xmath158 - ( 1-\epsilon'){r}\right ) } \right ) } \leq 1 \label{mbeta}\end{aligned}\ ] ] we know from that @xmath46 and @xmath105-achievable secrecy capacities satisfy the conditions @xmath159 |_{{r}=c_f^{\epsilon'}}$ ] and @xmath160 |_{{{r}}=c_f^{\epsilon}}=\expect[r_s({{\bf h}},p^{\ast})]$ ] , respectively . by lemma [ l : rb ] , we know that @xmath157 $ ] is a continuous function of @xmath39 , hence for any given @xmath103 , there exists an @xmath161 such that @xmath162 , and @xmath163=(1- \frac{\epsilon+\epsilon'}{2}){{r}}$ ] . furthermore , as @xmath164 , @xmath165 .
let us define a monotonically decreasing sequence @xmath166 , such that @xmath167 .
for any @xmath168 , find @xmath169 such that @xmath170 , and @xmath171=(1- \frac{\epsilon+\epsilon'(i)}{2}){r}(i)$ ] , therefore @xmath172 . from , we get @xmath173 since as @xmath174 , @xmath175 , @xmath176 and @xmath177 , where @xmath178 \\ & \leq \var[r_s({{\bf h}},p^{c_f^{\epsilon } } ) ] - c_f^{\epsilon}(1-\epsilon)\epsilon\end{aligned}\ ] ] the last inequality induces the upper bound , which concludes the proof .
in this section , we conduct simulations to illustrate our main results with two examples . in the first example , we analyze the relationship between @xmath46-achievable secrecy capacity and average power .
we assume that both the main channel and eavesdropper channel are characterized by rayleigh fading , where the main channel and eavesdropper channel power gains follow exponential distribution with means 2 and 1 , respectively .
since rayleigh channel is non - invertible , maintaining a non - zero secrecy rate with zero secrecy outage probability is impossible . in figure
[ fig : capacityresults ] , we plot the @xmath46-achievable secrecy capacity as a function of the average power , for @xmath179 outage probability , for both full csi and main csi cases .
it can be clearly observed from the figure that the gap between capacities under full csi and main csi vanishes as average power increases , which support the result of theorem [ t : highpowercapacity ] .
-achievable secrecy capacities as a function of average power , @xmath180,width=332 ] in the second example , we study the relationship between the buffer size , key loss ratio and the outage probability .
we assume that both the main and eavesdropper channel gains follow a chi - square distribution of degree 2 , but with means 2 and 1 , respectively .
we focus on the full csi case , and consider the scheme described in section [ s : finitebuffer ] .
we consider transmission at secrecy rate of @xmath39 with the use of the power allocation policy @xmath111 that solves the problem - . for @xmath181 , and the average power @xmath182 ,
we plot the key loss ratio , as a function of buffer size @xmath113 in figure [ fig : buffersizing1 ] , for @xmath183 , @xmath184 and @xmath185 , where @xmath102 is the @xmath46-achievable secrecy capacity .
it is shown in lemma [ t : losspr ] of section [ s : finitebuffer ] that expect the key loss ratio @xmath131 decreases as @xmath161 increases , which is observed in figure [ fig : buffersizing1 ] .
finally , we study the relationship between the secrecy outage probability and the buffer size for a given rate . in figure
[ fig : buffersizing2 ] , we plot the secrecy outage probabilities , denoted as @xmath105 , as a function of buffer size @xmath113 for the same encoder parameters . on the same graph
, we also plot our asymptotic result given in theorem [ eq : buffervsoutage ] , which provides an upper bound on the required buffer size to achieve @xmath105 outage probability for rate @xmath102 , with the assumption that is an equality for any @xmath105 .
we can see that , this theoretical result serves as an upper bound on the required buffer size when @xmath186 , which is the additional secrecy outages due to key buffer overflows , is very small .
another important observation from figures [ fig : buffersizing1 ] and [ fig : buffersizing2 ] is that , for a fixed buffer size , although the key loss ratio decreases as @xmath39 increases , secrecy outage probability increases .
this is due to the fact that key bits are pulled from the key queue at a faster rate , hence the decrease in the key loss ratio does not compensate for the increase of the rate that key bits are pulled from the key queue , therefore the required buffer size to achieve same @xmath105 is higher for larger values of @xmath39 . , and outage probability @xmath105 ] , and outage probability @xmath105 ]
this paper obtained sharp characterizations of the secrecy outage capacity of block flat fading channels under the assumption full and main csi at the transmitter . in the two cases ,
our achievability scheme relies on opportunistically exchanging private keys between the legitimate nodes and using them later to secure the delay sensitive information .
we further derive the optimal power control policy in each scenario revealing an interesting structure based by judicious time sharing between time sharing and the optimal strategy for the ergodic .
finally , we investigate the effect of key buffer overflow on the secrecy outage probability when the key buffer size is finite .
first , we prove the achievability . consider a fixed power allocation function @xmath188 .
let us fix @xmath189/(1-\epsilon)$ ] .
we show that for any @xmath47 , there exist some @xmath1 and @xmath0 large enough such that the constraints in and are satisfied , which implies that any @xmath190/(1-\epsilon)$ ] is an @xmath46-achievable secrecy rate .
the outage capacity is then found by maximizing @xmath72/(1-\epsilon)$ ] over the set @xmath191 of power allocation functions .
our scheme utilizes secret key buffers at both the transmitter and legitimate receiver .
then , + i ) at the end of every block @xmath3 , using privacy amplification , legitimate nodes ( transmitter and receiver ) generate @xmath192 bits of secret key from the transmitted signal in that particular block , and store it in their secret key buffers .
we denote the generated secret key at the transmitter as @xmath193 , and at the receiver as @xmath194 .
+ ii ) at every block @xmath3 , @xmath195 , the transmitter pulls @xmath196 bits from its secret key buffer to secure the outage constrained message of size @xmath197 , using vernam s one time pad .
the receiver uses the same key to correctly decode the message .
we denote the pulled key at the transmitter as @xmath198 , and at the receiver as @xmath199 . for simplicity in analysis
, we assume that keys generated at @xmath200th superblock are used only in the @xmath201th superblock .
this stage is skipped in the first super - block , and when encoder outage @xmath202 occurs , which is the union of the following events : * channel outage ( @xmath203 ) : channel is not suitable for reliable transmission at rate @xmath161 , i.e. , @xmath204 . *
key outage ( @xmath205 ) : there are not enough key bits in the key queue to secure @xmath25 , i.e. , @xmath206 * artificial outage ( @xmath207 ) : the transmitter declares outage , even though reliable secure transmission of @xmath25 is possible .
this is introduced to bound the probability of key outages , and is explained in the outage analysis . *
encoding : * + our random coding arguments rely on an ensemble of codebooks generated according to a zero mean gaussian distribution with variance @xmath21 . of channel gains , and using a different gaussian codebook for every partition @xcite . ]
+ 1 ) when @xmath202 does not occur , the message is secured with the secret key bits pulled from the key queue , using one time pad @xmath208 clearly , @xmath209 .
furthermore , let @xmath210 . to transmit the one time padded message @xmath211 , the encoder randomly and uniformly chooses @xmath212 among @xmath213 , and transmits to codeword @xmath30 indexed by @xmath214 over the channel .
+ 2 ) when @xmath202 occurs , @xmath25 is not transmitted .
let @xmath215 .
the encoder randomly and uniformly chooses @xmath216 among @xmath217 , and transmits to codeword @xmath30 indexed by @xmath216 over the channel .
+ the reason for transmitting @xmath212 and @xmath216 is to confuse the eavesdropper to the fullest extent in the privacy amplification process . +
* decoding :
* + 1 ) when @xmath218 does not occur , the receiver finds the jointly typical @xmath219 pair , where @xmath220 denotes the received signal at block @xmath3 .
then , using one - time pad , the receiver obtains @xmath221 .
+ 2 ) when @xmath218 occurs , the receiver finds the jointly typical @xmath222 .
define the error events @xmath223 independent of whether the event @xmath218 occurs or not , the encoding rate is equal to @xmath224 , which is below the supremum of achievable main channel rates . furthermore , each element of @xmath225 is independently drawn from gaussian distribution of mean @xmath89 and variance @xmath21 .
therefore , random coding arguments guarantee us that @xmath226 , @xmath227 such that @xmath228 , @xmath229 , @xmath230 and @xmath231 .
* privacy amplification : * at the end of every block @xmath3 , the transmitter and receiver generate secret key bits , by applying a universal hash function on the exchanged signals in that particular block .
first , we provide the definition of a universal hash function .
( @xcite ) a class @xmath232 of functions @xmath233 is universal , if for any @xmath234 in @xmath235 , the probability that @xmath236 is at most @xmath237 when @xmath238 is chosen at random from @xmath232 according to a uniform distribution .
[ l : privacyamp ] for any @xmath239 , there exists @xmath240 such that , @xmath241 , and for any block @xmath3 , * when @xmath218 does not occur , the transmitter and receiver can generate secret key bits @xmath242)$ ] and @xmath243)$ ] respectively , such that @xmath244 if the error event @xmath245 does not occur , and @xmath246 * similarly , when @xmath218 occurs , the transmitter and receiver can generate secret key bits @xmath247)$ ] and @xmath248)$ ] respectively , such that @xmath244 if the error event @xmath249 does not occur , and , are satisfied .
the proof follows follows the approach of @xcite , which applies privacy amplification to gaussian channels .
first , we introduce the information theoretic quantities required for the proof . for random variables
@xmath250 , define * renyi entropy of @xmath251 as @xmath252 $ ] * min - entropy as @xmath251 as @xmath253 .
* conditional min - entropy of @xmath251 given @xmath1 as @xmath254 * @xmath35-smooth min - entropy of @xmath251 as @xmath255 . without loss of generality ,
we drop the block index @xmath3 and @xmath39 , and focus on the first block @xmath256 , and assume the event @xmath257 does not occur .
let @xmath258 $ ] , with sample realization sequences denoted by @xmath259 .
let @xmath260 , where @xmath232 denotes a random universal hash function that maps @xmath261 to to an r - bit binary message @xmath262 .
then , it is clear that if error event @xmath263 does not occur , @xmath264 since @xmath265 , for any choice of @xmath232 . to show that the security constraints - are satisfied
, we cite the privacy amplification theorem , which is originally defined for discrete channels . for this purpose ,
we define a quantization function @xmath266 , with sensitivity parameter @xmath267 .
let @xmath268 denote the quantized version of @xmath269 .
where @xmath270 denotes realization sequences .
then , by theorem 3 of @xcite there exists a universal function @xmath232 such that in the following parts of the proof of lemma [ l : privacyamp ] for notational simplicity . ]
@xmath271 now , we relate this expression to the shannon entropy of the message , conditioned on eavesdropper s actual received signal . using the facts @xmath272 and @xmath273 , it is easy to show that @xmath274 then , due to the asymptotic relationship between continuous random variables and their quantized versions @xcite , there exists a quantization function @xmath266 such that @xmath275 is small enough , and @xmath276 are satisfied . to relate min - entropy to shannon entropy
, we use the result of theorem 1 of @xcite ; @xmath277 , @xmath278 a block length @xmath279 such that @xmath280 , @xmath281 now , we proceed as follows , @xmath282 where follows from , and the appropriate choice of @xmath279
. follows from the fact that @xmath283 forms a markov chain
. follows from the fact that @xmath284 , and similarly @xmath285 , which is the eavesdropper s maximum achievable rate . for the choice of @xmath286 , from , , and
the fact that @xmath287 , we get @xmath288 since there exists some @xmath289 such that forall @xmath290 , hence , for @xmath291 , the constraints , are satisfied .
the proof for the case where @xmath257 occurs is very similar , and is omitted .
* equivocation analysis : * secrecy outage probability can be bounded above as @xmath292 where the first equality follows from the definition of secrecy outage , and follows from the union bound .
now , we upper bound the first term . for the choice @xmath293 such that @xmath294 , the equivocation at every block @xmath3 in case of no encoder outage can be bounded as @xmath295 where we use the notation @xmath296 , and omit the index @xmath297 from @xmath298 .
notice that @xmath299 , and due to our encoder structure , @xmath300 is transmitted only in @xmath3th block , and similarly , @xmath198 is generated in @xmath200th superblock .
then , due to the memoryless property of the channel , @xmath301 , hence @xmath302 follows .
the first term of @xmath303 follows since @xmath300 is independent of @xmath304 , and the second term of @xmath303 follows since @xmath305 forms a markov chain .
@xmath306 follows due to @xmath307 .
@xmath308 follows since there is no encoder outage , hence key outage , and @xmath309 follows since @xmath198 is pulled from the key buffers , which contain the pool of key bits @xmath310 generated during superblock @xmath200 , and follows from .
then , @xmath311 now , we bound the encoder outage probability . by the union bound , @xmath312 since @xmath188 , due to the definition in , , @xmath313 @xmath314}{1-\epsilon}\right ) \leq \epsilon\end{aligned}\ ] ]
where in @xmath315 , we interchangeably use @xmath316 due to stationarity of @xmath60 . note that , the events @xmath317 indexed by @xmath3 are i.i.d . here
, we introduce i.i.d .
artificial outages @xmath318 such that @xmath319 this would help us bound the probability of key outage . for @xmath3 , @xmath195 @xmath320
< 0\right ) \label{appa : keyqueue}\end{aligned}\ ] ] note that , the expression in represents a random walk with expected drift @xmath321-\delta - { r}(1-\epsilon)$ ] . for @xmath322-\delta}{1-\epsilon}$ ] , @xmath323 , hence by the law of large numbers , @xmath324 such that @xmath325 , @xmath326 , @xmath50 .
therefore , due to union bound and , , , for the choice @xmath327 , @xmath328 , @xmath329 , which satisfies .
* error analysis : * for @xmath293 such that @xmath330 , @xmath328 , @xmath313 , @xmath195 , @xmath331 where the first term can be bounded as @xmath332 due to definition of @xmath245 , and the choice of @xmath0 .
similarly , the third term can be bounded as @xmath333 due to definition of @xmath334 , and the choice of @xmath0 .
the second term can be bounded as @xmath335 where @xmath302 follows from the fact that keys used in @xmath201th superblock are generated in @xmath200th superblock , and @xmath303 follows due to the definitions of @xmath245 and @xmath249 .
therefore , @xmath336 . finally , @xmath337 where the last inequality follows from the fact that @xmath338 .
this concludes the achievability .
now , we prove the converse .
consider a power allocation function @xmath21 , which satisfies the average power constraint @xmath339 let @xmath47 .
it follows from the converse proof of ergodic secrecy capacity @xcite , and law of large numbers that @xmath340 such that for every @xmath2 , @xmath341 , and @xmath342 , the time - average equivocation rate .
] is bounded as @xmath343 if @xmath39 is an @xmath46 achievable rate , then @xmath344 such that @xmath345 @xmath346 where @xmath302 follows directly from the definition of the event @xmath347 , and @xmath303 follows from applying the secrecy outage constraint , and the law of large numbers . from ,
, it follows that any @xmath46-achievable rate @xmath39 satisfies @xmath348 since secrecy outage probability has to be satisfied , for @xmath3 , @xmath195 channel outage probability also has to be satisfied , i.e. , @xmath349 , which implies @xmath350 since @xmath52 and @xmath351 are both deterministic functions of the power @xmath21 and instantaneous channel gains @xmath352 , it follows that the power allocation function that maximizes @xmath353 under the constraints , is a stationary function of instantaneous channel gains @xmath352 .
interchanging the notations @xmath354 , @xmath65 and @xmath66 , we can see that for any @xmath46 achievable secrecy rate , the constraints - are satisfied , which completes the proof .
the proof is very similar to the proof for full csi , hence we only point out the differences . for full csi ,
key generation occurs at the end of every _ block _ , using privacy amplification . due to lack of eavesdropper channel state at the legitimate nodes ,
this is no longer possible .
however , as shown in @xcite , it is still possible to generate secret key bits over a _
superblock_. the following lemma replaces lemma [ l : privacyamp ] in the full csi case .
let us define @xmath355 , where @xmath356,~{{{\cal o}}}_{\text{enc}}(s , b,{r})\mbox { does not occur}\\ [ w_{x2}(s , b)],~{{{\cal o}}}_{\text{enc}}(s , b,{r})\mbox { occurs } \end{cases } \end{aligned}\ ] ] and similarly define @xmath357 .
there exists @xmath358 such that , @xmath359 , and for any superblock @xmath201 , the transmitter and receiver can generate secret key bits @xmath360 and @xmath361 respectively , such that @xmath362 if none of the error events @xmath363 , @xmath364 occur , and @xmath365 - \delta ) \label{appb : privacyamp1}\\ \frac{1}{nb } i(v(s);{{\bf z}}^{sb},{{\bf h}}^{sb},g ) & \leq \delta
\label{appb : privacyamp2 } \end{aligned}\ ] ] the proof is very similar to the proof of lemma [ l : privacyamp ] , and is omitted here .
following the same equivocation and error analysis in the full csi case , we can see that any rate @xmath366/(1-\epsilon)$ ] is achievable .
the converse proof is also the same as in full csi case , and is omitted here .
the parameter @xmath367 is the maximum value for which the problem - has a solution ; hence the average power constraint is active .
moreover , the outage constraint is also active , and due to the fact that @xmath368 is a concave increasing function of @xmath60 , we have @xmath369 , since otherwise one can further increase @xmath367 to find a power allocation function that satisfies the equality . since for a given @xmath64 , the power allocation function that yields @xmath367 is @xmath370 , we have @xmath371 where @xmath372 the set of channel gains for which the system operates at rate @xmath367 , and @xmath373 . the set @xmath372 contains channel gains @xmath64 for which @xmath370 takes minimum values , so that the average power constraint is satisfied for the maximum possible @xmath161 . since @xmath374 is a decreasing function of @xmath375 , one can see that the choice of @xmath372 that yields @xmath367 is @xmath376 . since the probability density function of @xmath84 is well defined , @xmath377 , hence @xmath378 , which , along with @xmath23 , implies that @xmath379 .
let @xmath380 .
then , any @xmath60 that satisfies @xmath381 , would also satisfy @xmath382 .
so , the set of power allocation functions that satisfy shrinks as @xmath161 increases , hence @xmath157 $ ] is a non - increasing function of @xmath161 .
now , we prove that @xmath157 $ ] is continuous . from lemma
[ l : fullcsipower ] , we know that @xmath383 where @xmath384 and @xmath385 are constants that satisfy and with equality with respect to parameters @xmath39 and @xmath386 , respectively .
let us define another power allocation function @xmath387 such that @xmath388 it is easy to see that @xmath389 \leq \expect[r_s({{\bf h}},p^{{{r}}'})]$ ] . combining the facts i ) for any @xmath64 , @xmath390 is a continuous function of @xmath161 ii ) @xmath71 is a continuous function of @xmath391 iii ) integration preserves continuity , we can see that @xmath392 is a continuous function of @xmath393 .
hence , for any @xmath394 , one can find a @xmath47 such that for any @xmath395 , @xmath396 , @xmath397 - \expect[r_s({{\bf h}},p^{{{r } } ' } ) ] & \leq \expect[r_s({{\bf h}},p^{{{r}}})]- \expect[r_s({{\bf h}},\tilde{p}^{{{r } } ' } ) ] \\ & \leq \int r_s({{\bf h}},p_{\text{inv}}^{{r } } ) - r_s({{\bf h}},p_{\text{inv}}^{{{r } } ' } ) \ind({{\bf h}}\in { { \cal g}}(\lambda_{{r}},k_{{r } } ) ) f({{\bf h}})d{{\bf h}}\nonumber \\ & \leq \gamma \nonumber\end{aligned}\ ] ] which proves that @xmath398 $ ] is a left continuous function of @xmath161 .
following a similar approach , it can also be shown that @xmath157 $ ] is continuous from the right .
if @xmath157|_{{r}=0}=0 $ ] , then the unique solution of @xmath399/(1-\epsilon)$ ] is @xmath400 .
so , consider @xmath157|_{{r}=0}=0 $ ] .
it is easy to see that , @xmath401}{{{r}}_{\max}}\leq(1-\epsilon)$ ] , since @xmath402 & = \int_{h_m \geq c } r_s({{\bf h}},p ) f({{\bf h } } ) d{{\bf h}}\\ & \leq r_m({{\bf h}},p)(1-\epsilon ) \\ & = { { r}}_{\max } ( 1-\epsilon)\end{aligned}\ ] ] follows from definition of parameter @xmath99 , and the inequality @xmath403 . combining the facts that , the function @xmath404}{{{r}}}$ ] is continuous and strictly decreasing on @xmath405 $ ] , @xmath406}{{{r}}}= \infty$ ] and @xmath401}{{{r}}_{\max}}\leq(1-\epsilon)$ ] , by the intermediate value theorem , there exists a unique @xmath407 , which satisfies @xmath408/(1-\epsilon)$ ] .
we use lagrangian optimization approach to find @xmath111 .
we can express @xmath157 $ ] given in - as @xmath409 where the lagrangian @xmath410 is given by the equation . ]
@xmath411 \label{lagrangian2}\end{aligned}\ ] ] here , @xmath412 is a set which consists of @xmath64 for which @xmath413 must be satisfied
. we will show in this proof that it is of the form .
this problem is identical to , since their constraint sets are identical .
hence solution of this problem would also yield @xmath111 . in the following two - step approach ,
we proceed to find @xmath111 .
let us fix @xmath414 . 1 . for any @xmath415 , we find @xmath416 , which is defined as @xmath417 2 .
using the result of part 1 , we find @xmath111 , by finding the set @xmath412 that maximizes @xmath418 , subject to a constraint @xmath419 .
we start with step 1 .
since both @xmath420 and @xmath161 are fixed , therefore we drop them from @xmath421 and @xmath422 , in the following parts to simplify the notation .
[ appc : arbitraryglemma ] if the problem has a feasible solution , then it could be expressed as @xmath423^+\textbf{1}({{\bf h}}\in { { \cal g } } ) \label{objective1soln}\ ] ] where @xmath424 and @xmath425 are given in and , respectively
. we will interchangeably use @xmath426.$ ] due to , @xmath427 , @xmath428 .
hence , there is a minimum power constraint for set @xmath412 , as @xmath429 define @xmath372 as the set in which the minimum power constraint is not active , i.e. , @xmath430 where @xmath431 is complement of @xmath412 .
first , we focus on the solution in the nonboundary set .
since the optimal solution must satisfy the euler - lagrange equations , @xmath432 for @xmath433 , we get the following condition @xmath434 whose solution yields @xmath435\nonumber\end{aligned}\ ] ] if for some @xmath433 , the value @xmath60 is negative , then due to the concavity of @xmath436 with respect to @xmath60 , the optimal value of @xmath60 is zero @xcite .
therefore , the solution yields @xmath437 combining the result with the minimum power constraint inside set @xmath412 , the solution of yields , which concludes the proof .
now , we find @xmath111 .
we proceed by further simplifying the lagrangian in , for the case where @xmath438 , for a given @xmath412 as follows .
@xmath439f({{\bf h}})d{{\bf h}}\nonumber\\ & \quad + \int_{{{\bf h}}\notin { { \cal g}}}\left [ r_s({{\bf h}},p)-\lambda p({{\bf h } } ) \right]f({{\bf h}})d{{\bf h}}\nonumber\\ & = \quad \int \left [ r_s({{\bf h}},p_{\text{wf}})-\lambda p_{\text{wf}}({{\bf h } } ) \right]f({{\bf h}})d{{\bf h}}\nonumber\\ & \quad+ \int_{{{\cal g } } } \left\ { \left[r_s({{\bf h}},p_{\text{inv}})-r_s({{\bf h}},p_{\text{wf}})\right]^+ \right.\nonumber\\ & \qquad-\lambda \left.\left[p_{\text{inv}}({{\bf h}})-p_{\text{wf}}({{\bf h}})\right]^+\right\ } f({{\bf h}})d{{\bf h}}\label{jeqnfinal2 } \end{aligned}\ ] ] after this simplification , the first term in does not depend on @xmath412 .
we conclude the proof by showing that @xmath440 where the set @xmath441 is defined as follows , @xmath442^+ -\lambda \left[p_{\text{inv}}({{\bf h}})-p_{\text{wf}}({{\bf h}})\right]^+ \geq k \right\}\label{optimalg}\end{aligned}\ ] ] where the parameter @xmath443 is a constant that satisfies @xmath444
. we prove this by contradiction .
first define @xmath445^+ -\lambda \left[p_{\text{inv}}({{\bf h}})-p_{\text{wf}}({{\bf h}})\right]^+$ ] .
then , it follows from that @xmath441 is the set that maximize , so @xmath446 assume that some other @xmath447 is optimal , where @xmath448 .
however , we have @xmath449 since @xmath450 and @xmath451 by definition .
this contradicts our assumption that @xmath452 is optimal .
note that , @xmath441 is identical to .
this concludes the proof .
the proof goes along similar lines as in appendix [ a : fullcsipower ] , so we skip the details here .
we solve the problem for a fixed @xmath453 .
first , for any given @xmath454 , we define the following problem , the solution of which yields @xmath455 .
@xmath456,p ) \geq { { r } } , \forall h_m \in { { \cal g}}\label{appd : minrateconst } \end{aligned}\ ] ] [ appd : generalpowerlemma ] if the problem has a feasible solution , then it can be expressed as @xmath457 the proof uses the same approach as in proof of lemma [ appc : arbitraryglemma ] .
we define the set @xmath372 such that for any @xmath458 , the minimum rate constraint in is not active . since the optimal solution must satisfy the euler lagrange equations , we have @xmath459 if we solve the equation for any given @xmath375 , we get @xmath460 if the power allocation function that solves the equation is negative , then by the convexity of the objective function @xcite , the optimal value of @xmath62 is @xmath89 . hence , we get @xmath461 as the resulting power allocation function .
whenever the minimum rate constraint is active , we get the channel inversion power allocation function , @xmath390 . now , using lemma [ appd : generalpowerlemma ] , we solve the following problem , @xmath462 the solution of which yields @xmath96 .
lemma [ appd : generalpowerlemma ] proves that the solution is a time - sharing between @xmath463 and @xmath464 .
now , we find the optimal @xmath412 . the solution of is of the form , with the set @xmath465 , where @xmath99 is a constant which solves @xmath466 .
let @xmath467 and @xmath468 be the power allocation functions that are solutions of given the sets @xmath469 and @xmath452 , respectively .
we show that , any choice of @xmath447 , such that @xmath470 is suboptimal , i.e. , @xmath471 we continue as follows .
@xmath472,p_{\text{inv}})-r_s([h_m , h_e],p_{w})\right]^+ -\lambda \left[p_{\text{inv}}(h_m,{{r}})-p_{\text{wf}}(h_m,\lambda)\right]^+ \right ) f(h_m)dh_m \right\ } f(h_e ) dh_e \\ & - \int \left\ { \int_{h_m \in { { \cal g}}'}\left ( \left[r_s([h_m , h_e],p_{\text{inv}})-r_s([h_m , h_e],p_{w})\right]^+ - \lambda \left[p_{\text{inv}}(h_m,{{r}})-p_{\text{wf}}(h_m,\lambda)\right]^+ \right ) f(h_m)dh_m \right\ } f(h_e ) dh_e\end{aligned}\ ] ] note that , for any @xmath473 and @xmath474 , we have @xmath475 . since @xmath476 and @xmath477 , we have @xmath478^+ \leq \left[p_{\text{inv}}(h_m'',{{r}})-p_{w}(h_m'',\lambda)\right]^+\ ] ] since @xmath479 is a concave increasing function of @xmath63 @xcite , and for @xmath480 , we have @xmath481 .
therefore for any @xmath482 , we have @xmath483,p_{\text{inv}})-r_s([h_m',h_e],p_{w})\right]^+ -\lambda \left[p_{\text{inv}}(h_m',{{r}})-p_{w}(h_m',\lambda)\right]^+ \\
-&\left[r_s([h_m'',h_e],p_{\text{inv}})-r_s([h_m'',h_e],p_{w})\right]^+ + \lambda \left[p_{\text{inv}}(h_m'',{{r}})-p_{w}(h_m'',\lambda)\right]^+ \geq 0\end{aligned}\ ] ] combining this result with the packing arguments following in appendix [ a : fullcsipower ] , we get @xmath471 hence concluding the proof .
note that , this result can also be proved using the arguments of section 4 in @xcite .
due to theorem 1.2 of section vi in @xcite , it suffices to show that @xmath136 is a positive recurrent regenerative process .
note that @xmath136 is a markov process with an uncountable state space @xmath484 $ ] , since @xmath136 can be written as @xmath485 where @xmath115 and @xmath486 are i.i.d . ,
and @xmath487 depends only on @xmath136 and @xmath115 .
therefore , @xmath488 is independent of @xmath489 given @xmath136 , hence markovity follows .
now , we prove that @xmath136 is a recurrent regenerative process where regeneration occurs at times @xmath490 such that @xmath491 .
a sufficient condition for this is to show that @xmath136 has an accessible atom @xcite .
assume @xmath494 , @xmath495 $ ] .
note that , @xmath115 and @xmath496 are both i.i.d .
also note that , @xmath497 @xmath123 .
is sufficient to achieve @xmath105 secrecy outage probability . ]
find @xmath498 such that @xmath499 @xmath123 .
let @xmath500 .
then , @xmath501 since @xmath136 is a regenerative process , we know that @xmath502 are i.i.d
. random variables .
define a random variable @xmath503 , with distribution identical to @xmath504 .
now we show that @xmath136 is positive recurrent , by showing @xmath505 < \infty$ ] .
consider another recursion @xmath506 with @xmath507 .
it is clear that @xmath508 is also regenerative , where regeneration occurs at @xmath509 , where @xmath510 , and let @xmath511 be equal in distribution to @xmath512 .
note that @xmath508 is regenerative both at states @xmath89 and @xmath113 .
let @xmath520 $ ] denote the expected time for the process @xmath508 to hit @xmath89 from @xmath113 , and @xmath521 $ ] denote the expected time to hit @xmath113 from @xmath89 .
then , @xmath522 \leq \expect[\tau'_1 ] + \expect[\tau'_2 ] \label{tbar}\end{aligned}\ ] ] since the key queue has a negative drift , i.e. , @xmath523<0 $ ] , it is clear that @xmath520 < \infty$ ] .
now , we show that @xmath521 < \infty$ ] . following the approach of lemma [ l : atom ] ,
find @xmath394 such that @xmath524 @xmath123 .
let @xmath525 .
then , @xmath526 , and @xmath527&\leq \sum_{i=0}^{\infty}(\eta + i(\expect[\tau'_1]+\eta))\gamma^{\eta}(1-\gamma^{\eta})^i \\ & \leq \eta \gamma^{\eta}\sum_{i=0}^{\infty}(1-\gamma^{\eta})^i + \sum_{i=0}^{\infty}(1-\gamma^{\eta})^i i(\expect[\tau'_1]+\gamma^{\eta } ) \\ & < \infty\end{aligned}\ ] ] the first inequality follows from the fact that with probability @xmath528 , @xmath136 hits @xmath113 at @xmath529th block and with probability @xmath530 , key queue goes back to state @xmath89 at @xmath531+\gamma^{\eta})$]th block ( on average ) .
the last inequality follows from @xmath532 , and ratio test .
this result , along with and lemma [ tlemma ] concludes that @xmath136 is a positive recurrent regenerative process , which concludes the proof .
we follow an indirect approach to prove the lemma .
let @xmath533 denote the key queue dynamics of the same system for the infinite buffer case ( @xmath534 ) .
first , we use the heavy traffic results in @xcite to calculate the overflow probability of the infinite buffer queue .
then , we relate the overflow probability of infinite buffer system to the loss ratio of the finite buffer queue .
the dynamics of the infinite buffer queue is characterized by @xmath535 where @xmath536 .
the heavy traffic results we will use are for queues that have a stationary distribution . since it is not clear whether @xmath537 is stationary or not , we will upper bound @xmath537 by another stationary process @xmath538 , and the buffer overflow probability result we will get for @xmath538 will serve as an upper bound for @xmath537 . 1 .
if @xmath545 , then , using the facts @xmath546 and @xmath547 , we obtain @xmath548 which , using the described key queue recursions in , implies @xmath549 observe that , by ( [ eq : newqstar ] ) , @xmath550 which , in conjunction with ( [ eq : queuebound1 ] ) and @xmath551 , yields @xmath552 .
2 . if @xmath553 , then @xmath554 . we further consider two cases
. first , if @xmath555 , then , @xmath556 next , if @xmath557 , then @xmath558 which , combined with ( [ eq : queuebound3 ] ) , yields @xmath559 now , we show that @xmath538 converges in distribution to an almost surely finite random variable @xmath560 .
first , we need to show that the expected drift of @xmath538 is negative .
it is clear from that the expected drift of the process @xmath538 is equal to @xmath561-{r}(1-\epsilon)$ ] . from lemma [ l : rb ] in section [ s : fullcsipower ] , we know that @xmath157 $ ] is a non - increasing continuous function of @xmath161 .
therefore , @xmath565 it is a continuous function of @xmath161 .
furthermore , by definition of @xmath102 in , @xmath566 .
combining these two facts , we conclude that @xmath567 , for @xmath562 . combining lemma [ mulemma ] with the classic results by loynes @xcite
, we can see that @xmath538 converges in distribution to an almost surely finite random variable @xmath560 such that @xmath570 using , we finish the proof of the lemma . first , we prove that @xmath573 which is based on the heavy traffic limit for queues developed in @xcite , see also theorem 7.1 in @xcite . in order to prove , we only need to verify the following three conditions : i ) @xmath574 ; ii ) @xmath575 ; and iii ) the set @xmath576 of random variables indexed by @xmath161 is uniformly integrable .
\iii ) note that , @xmath161 lies on the interval @xmath580 $ ] , where @xmath367 , defined in lemma [ bmaxlemma ] then we have @xmath581 since @xmath71 is a continuous function of @xmath60 , and for any @xmath161 on the interval @xmath580 $ ] , @xmath582 , hence we can see that @xmath583 .
therefore , this class of random variables is uniformly integrable .
this completes the proof of .
this result , in conjunction with lemma [ l : stablelemma ] completes the proof . using lemma 1 in @xcite , we relate the loss ratio of our finite buffer queue @xmath136 to the overflow probability of the infinite buffer queue @xmath537 as follows @xmath397 \limsup_{t \to
\infty } l^t(m ) \leq \int_{x = m}^{\infty } \limsup_{t\to\infty}\pr(q(t)>x ) dx \label{lossoverflow}\end{aligned}\ ] ] combining lemma [ l : bufferoverflow ] with , the proof is complete .
u. maurer and s. wolf , `` information - theoretic key agreement : from weak to strong secrecy for free , '' _ advances in cryptology - euro - crypt 2000 , lecture notes in computer science 1807 _ , pp.351 - 368 , 2000 .
nascimento , j. barros , s. skludarek , and h. imai , `` the commitment capacity of the gaussian channel is infinite , '' _ ieee transactions on information theory _ , vol.54 , no.6 , pp.2785 - 2789 , june 2008 h. s. kim and n. b. shroff , `` on the asymptotic relationship between the overflow probability in an infinite queue and the loss probability in a finite queue , '' _ advances in applied probability _ ,
836 - 863 , dec . 2001 | this paper considers point to point secure communication over flat fading channels under an outage constraint .
more specifically , we extend the definition of outage capacity to account for the secrecy constraint and obtain sharp characterizations of the corresponding fundamental limits under two different assumptions on the transmitter csi ( channel state information ) .
first , we find the outage secrecy capacity assuming that the transmitter has perfect knowledge of the legitimate and eavesdropper channel gains . in this scenario , the capacity achieving scheme relies on opportunistically exchanging private keys between the legitimate nodes .
these keys are stored in a key buffer and later used to secure delay sensitive data using the vernam s one time pad technique .
we then extend our results to the more practical scenario where the transmitter is assumed to know only the legitimate channel gain . here ,
our achievability arguments rely on privacy amplification techniques to generate secret key bits . in the two cases
, we also characterize the optimal power control policies which , interestingly , turn out to be a judicious combination of channel inversion and the optimal ergodic strategy . finally , we analyze the effect of key buffer overflow on the overall outage probability . |
when particles attract by a short - range interaction with a large scattering length , their low - energy physics becomes _ universal _ , i.e. , independent of microscopic details @xcite .
ultracold atoms are ideal to study such universal physics because of the tunability of interatomic interactions and can provide insights applicable in a broad range of physics .
one of the most striking phenomena in universal systems is the efimov effect , i.e. , the formation of an infinite tower of three - body bound states characterized by discrete scale invariance @xcite .
although the efimov effect was originally predicted in the context of nuclear physics , it is now subject to extensive research in ultracold atoms @xcite . in this paper
, we propose novel systems in which _ two _ particles exhibit discrete scale invariance in their spectrum . in order for this to happen , their interaction needs to be scale invariant @xcite .
it is usually considered that the short - range interaction can be scale invariant only when the scattering length @xmath0 is set to be zero or infinite .
however , there is another possibility .
the scattering length is made space dependent and tuned to be proportional to the distance from the center of the system : . in this case , both the translational and rotational symmetries are partially broken .
while we expect similar physics , we shall not deal with such cases in this paper . ]
@xmath1 this interaction is _ scale invariant _ because there is no dimensional parameter in it .
the emergence of the discrete scale invariance for @xmath2 can be understood intuitively by using the born - oppenheimer approximation .
suppose one particle is much heavier than the other particle .
with the heavy particle fixed at @xmath3 , the light particle forms a bound state with binding energy @xmath4 $ ] , which in turn acts as an effective potential for the heavy particle .
therefore , one can design any attractive potential by tuning the space dependence of the scattering length . in particular ,
when @xmath5 , the effective potential becomes an inverse square potential for which it is well known that the spectrum exhibits discrete scale invariance @xcite . since
two particles are trapped about the center of the system with discrete scaling symmetry , we shall refer to our system as a _ scaling trap_.
this conclusion can be established for any mass ratio by solving the two - body problem exactly with the space - dependent scattering length .
while our idea works in any spatial dimensions , we shall give extensive and detailed analyses in one dimension ( secs .
[ sec : two - body][sec : eft ] ) and then present key results in two and three dimensions ( sec .
[ sec:2d-3d ] ) .
remarks on experimental realization with ultracold atoms are given in sec .
[ sec : remarks ] .
two interacting particles in one dimension are described by the schrdinger equation ( hereafter @xmath6 ) : @xmath7 \psi(x , x ) = e\,\psi(x , x).\ ] ] here , @xmath8 and @xmath9 are the total and reduced masses and @xmath10 and @xmath11 are the center - of - mass and relative coordinates , respectively . for a zero - range interaction whose strength depends on the position @xmath12 , the interaction potential is written as @xmath13 where @xmath14 is the space - dependent scattering length . for a bound - state solution with @xmath15 ,
the schrdinger equation is formally solved by @xmath16 where @xmath17 is the wave function in momentum space and @xmath18 is the fourier transform of the inverse scattering length . by integrating both sides of eq .
( [ eq : wave_function ] ) over @xmath19 , we obtain an integral equation solved by @xmath20 : @xmath21 we note that when @xmath22 is uniform , there is a single bound state with binding energy @xmath23 .
now for the linearly space - dependent scattering length @xmath24 , @xmath25 is ill defined because of the divergence at @xmath26 .
this divergence needs to be regularized , for example , by a sharp cutoff @xmath27 or by a smooth cutoff @xmath28 . in either case , the limit of an infinitesimal cutoff @xmath29 leads to @xmath30 for which the analytic solution of the integral equation ( [ eq : integral_eq ] ) is found to be [ eq : solution ] @xmath31}{\sqrt{(p/\kappa)^2 + 1 } } , \\[8pt ] \tilde\chi_-(p ) & = n_-\frac{\sin[s_-{\mathrm{arcsinh}}(p/\kappa)]}{i\sqrt{(p/\kappa)^2 + 1}}. \end{aligned}\ ] ] the @xmath32 ( @xmath33 ) sign corresponds to the even ( odd ) parity , and @xmath34 is the normalization constant to ensure @xmath35 and @xmath36 solves @xmath37 details of deriving eqs .
( [ eq : solution ] ) and ( [ eq : s_1d ] ) are presented in the appendix .
note that @xmath38 has a solution for any @xmath2 while @xmath39 has a solution only for @xmath40 .
the latter range of @xmath41 is assumed below unless otherwise stated .
( lower curve ) and @xmath42 ( upper curve ) determined by eq .
( [ eq : s_1d ] ) as functions of the slope of the scattering length @xmath43 .
[ fig : scale - factor_1d ] ] then the inverse fourier transform of @xmath44 leads to the wave function with two particles at the same point , @xmath45 , where [ eq : wave_function2 ] @xmath46 \psi_-(x,0 ) & = \kappa\,n_-\frac{\sinh\frac{\pi s_-}2}{\pi } k_{is_-}(\kappa|x|)\ \mathrm{sgn}(x ) . \end{aligned}\ ] ] an important observation is that this wave function toward the origin @xmath47 oscillates as @xmath48 $ ] .
the phase of this oscillation is fixed by the precise behavior of @xmath14 near @xmath26 , which is not universal and depends on experimental setups .
however , what is universal is that because of the logarithmic periodicity in @xmath49 , if @xmath50 is a solution , then @xmath51 are all solutions
. therefore , in each parity channel , there exists an infinite tower of two - body bound states characterized by discrete scale invariance : @xmath52 this is exactly the same physics as the efimov effect while the difference should be emphasized that our bound state consisting of two particles is trapped about the center of the system ( see fig .
[ fig : density ] below ) and the scale factor @xmath53 is tunable by the slope of the scattering length as shown in fig .
[ fig : scale - factor_1d ] .
we also note that the full scale invariance demonstrated by the classical hamiltonian ( [ eq : schrodinger ] ) with @xmath54 is broken down to its discrete subset by the scale @xmath55 generated in quantum mechanics .
this is known as a _
quantum anomaly _ @xcite .
we now determine single - particle density and momentum distributions in the two - particle trapped state with binding energy @xmath56 . with the use of the wave function obtained from eqs .
( [ eq : wave_function])([eq : solution ] ) , @xmath57 the density distribution of a particle with mass @xmath58 , @xmath59 , and its momentum distribution @xmath60 are plotted in figs .
[ fig : density ] and [ fig : momentum ] , respectively , for @xmath61 with equal masses @xmath62 .
in particular , the momentum distribution at @xmath63 has an oscillatory large momentum tail @xmath64 given by @xmath65 & \quad \times \frac{(1-is_\pm)\cosh\frac{\pi s_\pm}2+s_\pm\sinh\frac{\pi s_\pm}2 } { 2\cosh\pi s_\pm}\biggr]\biggr\}. \end{split}\ ] ] this logarithmic oscillation signals the discrete scale invariance , and exactly the same tail emerges in any few - body or many - body state as we will show later .
in contrast to the efimov effect in which the oscillatory tail appears at the subleading order @xcite , it appears at the leading order in our scaling trap .
this will make its observation easier by a time - of - flight measurement in ultracold - atom experiments .
a longstanding problem in efimov physics is whether the discrete scale invariance demonstrated for three particles persists for more particles @xcite .
it has been established that this is the case for four bosons @xcite while the situation is less clear for larger numbers of particles . here , we elucidate that the scaling trap realizes a novel pattern of discrete - scaling - symmetry violation for bosons .
different particle sectors obey different scaling laws , and incommensurate scalings among them result in the breakdown of discrete scale invariance .
this pattern can be explained easily in the limit @xmath66 where the born - oppenheimer approximation is applicable .
recall that one - dimensional bosons with a zero - range attraction form an @xmath67-body bound state with binding energy @xmath68 @xcite .
for the linearly space - dependent scattering length @xmath69 , this binding energy acts as an effective potential for the center - of - mass motion of the @xmath67-body cluster , which is described by @xmath70 \psi_n(x ) = e\,\psi_n(x).\ ] ] since this is the inverse square potential , @xmath67 bosons form an infinite tower of trapped states characterized by a discrete scaling symmetry set by a scale factor @xmath71 with @xmath72 however , these @xmath67-body trapped states for @xmath73 are actually unstable resonances coupled with continuum states .
for example , when @xmath74 , there are continuum states composed of a free particle and a two - body trapped state obeying a discrete scaling symmetry set by @xmath75
. because dilatations with respect to @xmath76 and @xmath77 are incommensurate , the discrete scale invariance breaks down for three bosons and hence more .
further insights into the nature of @xmath67-boson resonances can be developed in the same limit @xmath66 .
because the local binding energy scales as @xmath78 , their relative wave function is localized within a separation @xmath79 . on the other hand ,
their center - of - mass wave function is @xmath80 , which oscillates rapidly at @xmath81 and decays exponentially at @xmath82 .
this @xmath67-body cluster can decay into a deeper @xmath83-body cluster with @xmath84 , and its wave function also oscillates rapidly but with a different logarithmic period @xmath85 .
because of the resulting small overlap between their wave functions , the @xmath67-boson resonances are expected to have small decay widths and thus obey an _ approximate _
discrete scaling law set by @xmath86 .
-boson resonance energy @xmath87 in units of @xmath88 and unit lattice spacing versus excitation number @xmath89 at @xmath90 .
[ fig : resonances ] ] to confirm this consideration , we numerically computed @xmath67-boson resonance energies for @xmath91 . here , a hamiltonian lattice formalism and iterative eigenvector methods were used in a semi - infinite system with a hard - wall boundary at the origin .
where the born - oppenheimer approximation is applicable . ] at @xmath90 , decay widths of resonances are indeed found to be negligible , and scaling behaviors of resonance energies are clearly seen in fig . [
fig : resonances ] as linear behaviors for large @xmath89 in the logarithmic plot .
their scale factors are extracted as @xmath92 , @xmath93 , and @xmath94 , which are in good agreement with the born - oppenheimer approximation @xmath95 for @xmath91 , respectively .
in contrast to bosons , we did not observe any resonances in the spectrum of three fermions ( two of one component and one of the other ) with equal masses .
this is indeed expected because one - dimensional two - component fermions with equal masses and with a zero - range attraction do not form any bound states with more than two fermions @xcite .
therefore , their discrete scaling symmetry can not be violated by the pattern elucidated above . .
[ fig : phase_diagram ] ] on the other hand , there exists another pattern of discrete - scaling - symmetry violation which is common to bosons and fermions . for more than two particles , both even- and odd - parity
two - particle states contribute in general .
because they obey incommensurate scalings set by @xmath96 and @xmath97 , the discrete scale invariance breaks down .
however , this pattern does not take place in the range @xmath98 where only the even - parity channel exhibits the discrete scale invariance , and the odd - parity channel remains scale invariant [ see eq .
( [ eq : s_1d ] ) ] . here , it is possible that an arbitrary number of fermions maintains the discrete scaling symmetry set by @xmath96 . furthermore , when @xmath99 , both even- and odd - parity channels remain scale invariant and hence the whole system .
accordingly , as shown in fig .
[ fig : phase_diagram ] , we expect two - component fermions with equal masses to exhibit three phases with distinct symmetries as a function of the inverse slope : phases with full scale invariance at @xmath100 , discrete scale invariance at @xmath101 , and no scale invariance at @xmath102 . their many - body physics and phase transitions in between will be extremely interesting and should be explored in the future .
the scaling trap can be formulated in the language of effective field theories .
a local - field theory to be considered is [ eq : field_theory ] @xmath103 \\[6pt ] \label{eq : counterterm } & + \frac{g_+}{\mu}\psi_1^{\dagger}\psi_2^{\dagger}\psi_2\psi_1(0 ) + \frac{g_-}{\mu}\nabla[\psi_1^{\dagger}\psi_2^{\dagger}]\nabla[\psi_2\psi_1](0 ) , \end{aligned}\ ] ] where @xmath104 is summed and the space argument @xmath105 acts on all operators on its left .
the first part [ eq . ( [ eq : hamiltonian ] ) ] describes particles interacting by the zero - range interaction with the linearly space - dependent scattering length . as we discussed above , such an interaction is singular at the origin and needs to be regularized by introducing a cutoff .
the independence of physical quantities from the cutoff is ensured by counterterms . according to the above argument of discrete - scaling - symmetry violation , @xmath67-body counterterms , @xmath106 and @xmath107\nabla[\psi^n](0)$ ] ,
are needed for bosons each with @xmath108 . on the other hand , the two counterterms in eq .
( [ eq : counterterm ] ) are expected to be sufficient for an arbitrary number of two - component fermions with equal masses .
an expression for the cutoff - dependent coupling @xmath109 is obtained in the same spirit of ref .
we employ a sharp momentum cutoff @xmath110 and require that two - body physics becomes independent from the choice of @xmath111 .
the two - body sector of our field theory ( [ eq : field_theory ] ) is equivalent to the schrdinger equation in eq .
( [ eq : schrodinger ] ) but with the modified interaction potential @xmath112\frac{\delta(x)}{\mu}.\ ] ] accordingly , the integral equation in eq .
( [ eq : integral_eq ] ) is modified to @xmath113 & \times \left[\frac2c\ln\!\left(\frac{|p - p'|}\lambda\right ) + g_+(\lambda ) + g_-(\lambda)pp'\right]\tilde\chi(p ' ) .
\end{split}\ ] ] then we require that its solution @xmath44 for @xmath114 found in eq .
( [ eq : solution ] ) does not change when the cutoff is changed from @xmath111 to @xmath115 .
this requirement is satisfied by choosing [ eq : rg ] @xmath116}{s_+}\right\ } , \\[8pt ] \lambda^2g_-(\lambda ) & = \frac2c\,\alpha_- \frac{s_-\cos[\phi_-(\lambda ) ] + \sin[\phi_-(\lambda ) ] } { s_-\cos[\phi_-(\lambda ) ] - \sin[\phi_-(\lambda ) ] } , \end{aligned}\ ] ] where @xmath117 . with numerical constants @xmath118 and @xmath119 , the analytic expression ( [ eq : rg ] ) fits to numerical solutions of the integral equation ( [ eq : integral_eq2 ] ) accurately as shown in fig .
[ fig : limit_cycle ] . and
@xmath120 for @xmath121 , respectively . ] an important observation is that these couplings run in logarithmically periodic ways in the large cutoff limit @xmath122 .
this is known as a _ renormalization group limit cycle _ , and the efimov effect is its rare manifestation in physics @xcite . our scaling trap is newly added to a short list of systems demonstrating the limit cycle .
the field - theoretical formulation ( [ eq : field_theory ] ) is useful to derive universal relationships valid for any few - body or many - body state @xcite .
an operator product expansion of @xmath123 & = w_+(x_1,p_1){\mathcal{o}}_+(x_1 ) + w_-(x_1,p_1){\mathcal{o}}_-(x_1 ) + \cdots \notag\end{aligned}\ ] ] at @xmath124 is dominated by the following two local operators : @xmath125 { \mathcal{o}}_-(x ) & \equiv \nabla[\psi_1^{\dagger}\psi_2^{\dagger}]\nabla[\psi_2\psi_1](x ) .
\end{aligned}\ ] ] by matching the matrix elements of both sides of eq .
( [ eq : ope ] ) with respect to the two - particle trapped state obtained in sec .
[ sec : two - body ] , the wilson coefficient of @xmath126 is found to be @xmath127 accordingly , the momentum distribution of a particle with mass @xmath58 exhibits the oscillatory large momentum tail @xmath128 & \to t_+(p_1)\,{\mathcal{c}}_++t_-(p_1)\,{\mathcal{c}}_-\end{aligned}\ ] ] for any state of the scaling trap .
the functional form of each term is fixed by the two - body physics @xmath129 obtained in eq .
( [ eq : tail ] ) for equal masses while its overall magnitude is set by an analog of the contact density defined by @xmath130 by applying the hellmann - feynman theorem to the hamiltonian ( [ eq : field_theory ] ) , we find that @xmath131 measures how an energy of the state under consideration changes with respect to @xmath55 : @xmath132 in particular , the contact density is given by @xmath133 for a two - particle trapped state with binding energy in eq .
( [ eq : spectrum ] ) .
another universal relationship involving @xmath131 can be derived by considering a `` boundary '' operator product expansion of @xmath134 toward the origin @xmath47 . by matching the matrix elements of both sides of eq .
( [ eq : bope ] ) with respect to the two - particle trapped state obtained in sec .
[ sec : two - body ] , the wilson coefficient of @xmath126 is found to be @xmath135 where @xmath136 is obtained from the two - body wave function in eq .
( [ eq : wave_function2 ] ) : @xmath137\biggr|^2 , \end{split } \\[8pt ] \begin{split } u_-(x ) & = \biggl|n_-\frac{\sinh\frac{\pi s_-}2}{\pi}\gamma(is_- ) \\ & \quad \times \cos\!\left[s_-\ln\!\left(\frac{\kappa_-|x|}2\right ) - \arg\gamma(is_-)\right]\biggr|^2 .
\end{split } \end{aligned}\ ] ] therefore , the probability of finding two particles at the same point oscillates toward the origin as @xmath138 these universal relationships found in eqs .
( [ eq : momentum ] ) , ( [ eq : energy ] ) , and ( [ eq : pair ] ) are valid for any few - body or many - body state of the scaling trap .
so far we have focused on physics in one dimension , but our idea works equally well in two and three dimensions .
a two - body bound - state problem by a zero - range interaction in an arbitrary spatial dimension @xmath139 reduces to solving an integral equation which is an analog of eq .
( [ eq : integral_eq ] ) in real space : @xmath140^{2-d}}.\ ] ] here , @xmath141 is the binding energy , and @xmath142}\ ] ] and @xmath143\ ] ] are associated with wave functions with two particles at the same point .
the scattering length @xmath144 is defined so that when @xmath145 is uniform , there is a single bound state with binding energy @xmath23 .
is different from the conventional one defined by @xmath146 for which the binding energy is given by @xmath147 .
therefore , @xmath148 in this paper , where @xmath149 is the euler - mascheroni constant . ] the emergence of a scaling trap for the linearly space - dependent scattering length @xmath150 is deduced from the existence of solutions that oscillate as @xmath151 toward the origin @xmath152 . by substituting a scaling ansatz @xmath153 or @xmath154 into the integral equation ( [ eq : integral_eq_real ] ) with @xmath155 , we find that @xmath156 solves @xmath157 in @xmath158 with @xmath159 or @xmath160 in @xmath161
. for each angular momentum @xmath162 where @xmath156 has a solution , has a solution when @xmath163\approx7.124,\,1.781,\,0.9642,\,0.6552,\,\ldots$ ] in @xmath158 or when @xmath164 ^ 2\approx3.142,\,1.273,\,0.7854,\,0.5659,\,\ldots$ ] in @xmath161 for @xmath165 , respectively .
] there exists an infinite tower of two - particle trapped states characterized by discrete scale invariance : @xmath166 its scale factor @xmath167 is plotted in fig .
[ fig : scale - factor_2d-3d ] .
we expect similar patterns of discrete - scaling - symmetry violation for more than two particles as elucidated in one dimension .
in this paper , we proposed scaling traps in which two particles form an infinite tower of trapped states characterized by discrete scale invariance .
the key idea is to make the scattering length proportional to the distance from the center of the system so that a short - range interaction becomes scale invariant . because of the lack of translational symmetries , center - of - mass and relative motions of the pair are coupled , and its relative motion generates an effective inverse square potential for its center - of - mass motion .
such a space - dependent interaction can be realized in ultracold - atom experiments by a magnetic- or optical - field - induced feshbach resonance with spatially varying magnetic- or optical - field intensity @xcite or , if the system is confined in lower dimensions , by a confinement - induced resonance with transverse confinement lengths varying along a longitudinal direction @xcite .
if two particles correspond to different spin states of a fermionic atom , our two - body bound states with efimov character are long - lived because three - body recombinations are strongly suppressed by the pauli exclusion principle @xcite .
furthermore , a scale factor can be easily controlled by the slope of the scattering length .
this will greatly facilitate an observation of the discrete scale invariance , for example , by a radio - frequency spectroscopy or a time - of - flight measurement .
therefore , the scaling trap overcomes common difficulties in ultracold - atom experiments of efimov physics arising from an instability of three - body bound states and their sizable scale factor @xmath168 .
advantages of the scaling trap over the efimov effect are summarized in table [ tab : comparison ] .
.comparison between the efimov effect and the scaling trap in terms of ( i ) the dimensionality in which they can appear , ( ii ) the required number of particles , ( iii ) the stability of the bound states , ( iv ) the order at which the large momentum tail oscillates , and ( v ) the tunability of the scale factor @xmath169 .
[ tab : comparison ] [ cols="^,^,^,^,^,^,^,^,^",options="header " , ] we also elucidated that the discrete scaling symmetry emergent for two particles is inevitably violated for three or more bosons due to the appearance of resonance states obeying different scaling laws while they are absent for fermions .
it is possible that insights developed here shed new light on efimov physics .
our scaling trap thus serves as a tunable model system to investigate universal physics involving scale invariance , quantum anomaly , and renormalization group limit cycle , which are important in a broad range of quantum physics .
this work was supported by a lanl oppenheimer fellowship and the us department of energy under contract no .
de - fg02 - 03er41260 .
the integral equation ( [ eq : integral_eq ] ) in the limit of an infinitesimal cutoff @xmath29 becomes @xmath170 the analytic solution to this integral equation presented in eqs .
( [ eq : solution ] ) and ( [ eq : s_1d ] ) can be derived in a way similar to that used in ref .
@xcite . in the even - parity channel
@xmath171/2 $ ] , the integral equation ( [ eq : integral_eq3 ] ) in terms of a new variable @xmath172 can be written as @xmath173 \notag & = -\frac2c\sqrt\frac{m}{\mu } \int\!\frac{dq'}{2\pi}\ln[\epsilon\kappa|\sinh(q - q')|]\cosh q'\tilde\chi_+(q').\end{aligned}\ ] ] by defining the fourier transform of the integral kernel by @xmath174 the integral equation ( [ eq : integral_even ] ) can be brought into a differential equation : @xmath175 since @xmath176 is an even function , its solution is easily found to be @xmath177 with @xmath178 satisfying @xmath179 similarly , in the odd - parity channel @xmath180/2 $ ] , the integral equation ( [ eq : integral_eq3 ] ) in terms of a new variable @xmath172 can be written as @xmath181 \notag & = -\frac2c\sqrt\frac{m}{\mu } \int\!\frac{dq'}{2\pi}\ln\left|\tanh\frac{q - q'}2\right|\cosh q'\tilde\chi_-(q').\end{aligned}\ ] ] by defining the fourier transform of the integral kernel by @xmath182 the integral equation ( [ eq : integral_odd ] ) can be brought into a differential equation : @xmath183 since @xmath184 is an odd function , its solution is easily found to be @xmath185 with @xmath39 satisfying @xmath186 e. braaten and h .- w .
hammer , _ universality in few - body systems with large scattering length _ , phys .
428 * , 259 ( 2006 ) .
v. efimov , _ energy levels arising from resonant two - body forces in a three - body system _ , phys . lett .
b * 33 * , 563 ( 1970 ) .
y. castin and f. werner , _ single - particle momentum distribution of an efimov trimer _ ,
a * 83 * , 063614 ( 2011 ) .
e. braaten , d. kang , and l. platter , _ universal relations for identical bosons from three - body physics _ ,
lett . * 106 * , 153005 ( 2011 ) .
a. deltuva , _ properties of universal bosonic tetramers _ , few - body syst .
( to be published ) [ arxiv:1202.0167 ( physics.atom-ph ) ] , and references therein .
f. bedaque , h .- w . hammer , and u. van kolck , _ renormalization of the three - body system with short range interactions _ , phys .
lett . * 82 * , 463 ( 1999 ) ; _ the three boson system with short range interactions _ , nucl .
a * 646 * , 444 ( 1999 ) .
k. g. wilson , _ the origins of lattice gauge theory _
b ( proc . suppl . )
* 140 * , 3 ( 2005 ) . c. chin , r. grimm , p. julienne , and e. tiesinga , _ feshbach resonances in ultracold gases _ , rev . mod . phys . *
82 * , 1225 ( 2010 ) .
m. olshanii , _ atomic scattering in the presence of an external confinement and a gas of impenetrable bosons _ , phys .
lett . * 81 * , 938 ( 1998 ) .
d. s. petrov and g. v. shlyapnikov , _ interatomic collisions in a tightly confined bose gas _ , phys .
rev . a * 64 * , 012706 ( 2001 ) .
d. s. petrov , c. salomon , and g. v. shlyapnikov , _ weakly bound dimers of fermionic atoms _ , phys . rev .
* 93 * , 090404 ( 2004 ) ; _ scattering properties of weakly bound dimers of fermionic atoms _ , phys .
a * 71 * , 012708 ( 2005 ) .
a. o. gogolin , c. mora , and r. egger , _ analytical solution of the bosonic three - body problem _ ,
lett . * 100 * , 140404 ( 2008 ) . | when the scattering length is proportional to the distance from the center of the system , two particles are shown to be trapped about the center .
furthermore , their spectrum exhibits discrete scale invariance , whose scale factor is controlled by the slope of the scattering length .
while this resembles the efimov effect , our system has a number of advantages when realized with ultracold atoms .
we also elucidate how the emergent discrete scaling symmetry is violated for more than two bosons , which may shed new light on efimov physics .
our system thus serves as a tunable model system to investigate universal physics involving scale invariance , quantum anomaly , and renormalization group limit cycle , which are important in a broad range of quantum physics . |
query complexity theory is mainly concerned with the computational cost required to determine some specific property of functions .
the cost is measured by the number of queries which can be addressed to a black box " device ( the oracle ) that outputs instantaneously an answer to a query .
various results concerning query complexity of boolean functions , i.e. functions with a finite - size domain which have two possible output values only , exist in the literature @xcite .
the first of this sort of problems which has been addressed by the quantum computational theory , is represented by the deutsch - josza algorithm @xcite .
it involves the decision whether a given boolean function of @xmath0 binary digits is _ constant _ ( that is whether it outputs the same binary digit for every input value ) or _ balanced _ ( that is whether it outputs @xmath2 on exactly half inputs and @xmath3 on the other half ) and it can be solved with certainty by querying a quantum oracle just once . this task is successfully achieved by exploiting two essential features of quantum mechanics : its linearity , which allows the simultaneous evaluation of a function on linear superpositions of its input values , and the interference between amplitudes , which raises the probability of obtaining the desired results . the _ a - priori _ knowledge that the function computed by the quantum oracle is either constant or balanced , is crucial in solving the problem exactly , i.e. with zero probability of error . in this paper
, we address the question of deciding the constancy of a generic function chosen at random among the set of @xmath4 @xmath0-bit boolean functions .
in such a case , the iteration of the same algorithmic procedure permits one to decide about the constancy of such generic function with a probability of error which depends on @xmath0 and on the number @xmath1 of quantum oracle - queries .
unfortunately , this probability tends to @xmath2 in the worst case scenario , namely when the given function outputs the same value for every input but one .
nevertheless , it is our purpose to determine the efficiency of the quantum algorithm and the range of values of @xmath0 and @xmath1 for which it is more efficient compared to the unique classical procedure which involves querying the oracle successively until two different outputs are encountered .
not too surprisingly , due to classical conditioning of each output on the previous ones , in some specific situations ( the worst cases ) , for large number of inputs and nearly as many oracle - queries , the efficiency of the classical algorithm turns out to overcome that of the quantum one .
moreover , in the average case , i.e. when the @xmath0-bit boolean function is sorted at random , the quantum algorithm is always preferable to the classical one .
the general problem we are going to tackle is as follows .
suppose that we are given an oracle which instantaneously computes an arbitrary n - bit boolean function @xmath5 on inputs that , as usual , will be enumerated by the integers @xmath6 .
our goal is to devise a quantum algorithm , involving oracle queries , in order to determine , with some probability of error , whether such a function is constant or not .
since , by hypothesis , we have no _ a - priori _ information or knowledge about @xmath7 , we can only suppose that it has been sorted completely at random within the set @xmath8 of @xmath4 n - bit boolean functions .
if we have at our disposal a classical oracle which outputs the value of @xmath7 , and which can be queried at most @xmath1 times , the only conceivable classical algorithm able to solve the problem is@xmath9 : _ query the oracle @xmath1 times or until two output values are different : in the latter case stop querying and declare that the function is not constant , otherwise say that @xmath7 is constant _ " . of course , this simple algorithm fails to give the right answer only when the function is actually non - constant and when we have obtained @xmath1 consecutive equal digits in the querying procedure .
in fact , the function @xmath7 might output a different digit on any of the untried inputs , so that one has to evaluate the probability of such an event .
we start by considering the worst possible scenario , that is the unknown function @xmath7 , drawn from the set @xmath8 , actually outputs the same bit , say @xmath3 , for all of its inputs but one ( for simplicity of notation we will say that such a function is of @xmath10-type ) .
the probability @xmath11 of declaring erroneously constant such a function , coincides with the probability of obtaining @xmath1 consecutive digits @xmath3 after querying the oracle @xmath1 times : @xmath12 = \textrm{prob}[0 ] \cdot \big ( \prod_{\ell=1}^{k-1}\textrm{prob}[\,0 | \underbrace{0\dots 0\,}_{\ell } ] \big),\ ] ] where we have indicated with @xmath13 $ ] , @xmath14 , the conditional probability of getting @xmath3 after the @xmath15-th oracle query if a sequence of @xmath16 consecutive zeroes has already been obtained .
since the function @xmath7 is of @xmath10-type , one easily calculates @xmath17=(2^{n}-1)/2^{n}$ ] and @xmath18= ( 2^{n}-\ell-1)/(2^{n}-\ell)$ ] , whence @xmath19 this probability distribution is equal to zero for @xmath20 ; indeed , due to statistical correlations between subsequent outputs , in such a case we can determine with certainty if the function is constant or not .
as we will see , this is not the case with the deutsch - josza quantum algorithm which , unless the function is either constant or balanced , is not able to give the correct answer with certainty .
first , let us briefly recall the basics of that algorithm whose associated quantum circuit is the following : ( 500,70 ) ( 120,40)(30,20 ) ( 220,40)(30,20 ) ( 120,0)(20,20 ) ( 170,0)(30,60 ) ( 110,10)(10,0)10 ( 140,10)(20,0)30 ( 200,10)(20,0)60 ( 110,50)(10,0)10 ( 150,50)(60,0)20 ( 200,50)(10,0)20 ( 250,50)(10,0)10 ( 225,47)@xmath21 ( 125,47)@xmath21 ( 178,27)@xmath22 ( 125,7)@xmath23 ( 90,7)@xmath24 ( 80,47)@xmath25 ( 270,47)@xmath26 we have indicated with @xmath27 the @xmath0-fold hadamard transformation on a @xmath0-qubit state and with @xmath28 the quantum oracle associated to @xmath7 .
it is defined by @xmath29 , where @xmath30 denotes the sum modulus @xmath31 , @xmath32 is an @xmath0-qubit state such that @xmath33 , and @xmath34 is a one - qubit state . if the initial input states are those indicated in the figure , it is possible to prove @xcite that the @xmath0-qubit output state @xmath35 in the figure above , turns out to be@xmath36 : @xmath37 where @xmath38 is the bitwise inner product of @xmath39 and @xmath40 , i.e. @xmath41 .
therefore , according to eq .
( [ worst3 ] ) , a measurement process performed onto the @xmath0-qubit computational basis @xmath42 yields the result @xmath43 with the following probability@xmath36 : @xmath44 = \big\vert \frac{1}{2^{n } } \sum_{x=0}^{2^{n}-1 } ( -1)^{f(x)}\big\vert^{2}\:.\ ] ] if the function @xmath7 is constant , the result @xmath43 is obtained with certainty ; this in turn implies that , if any other value @xmath45 is obtained , the function can not be constant . yet , if the function is of @xmath10-type , the result @xmath43 is obtained with a probability strictly smaller than @xmath2 , given by : @xmath46 = \big\vert \frac{1}{2^{n } } \sum_{x=0}^{2^{n}-1 } ( -1)^{f(x)}\big\vert^{2}= \big\vert \frac{1}{2^{n } } ( -2 + 2^{n})\big\vert^{2}= \big ( 1-\frac{1}{2^{n-1 } } \big)^{2}\:.\ ] ] the simplest quantum procedure to determine whether the function @xmath7 is constant or not , consists of a simple iteration of the quantum circuit described above and it amounts to :
_ iterate the deutsch - jozsa algorithm @xmath1 times or until a measurement result @xmath45 appears : in the latter case stop querying and declare that the function is not constant , otherwise say that @xmath7 is constant _ " .
we denote by @xmath47 the probability that , after obtaining @xmath43 in @xmath1 consecutive measurements , we are wrong in asserting the constancy of the @xmath10-type function . according to eq .
( [ worst4.5 ] ) , one gets : @xmath48 it is worth noting that , contrary to the classical case where the probability of obtaining the same output value changes , by conditioning , after each oracle query , the consecutive ( quantum ) oracle - queries are independent from each other .
therefore , the probability of obtaining @xmath43 , @xmath1 times consecutively , is the product of the probabilities of the uncorrelated events ; quantum conditioning on past results would certainly improve the efficiency of the algorithm , however we do not know of any such technique . in order to compare the quantum querying procedure with respect to the classical one in this worst case , we define the following efficiency function @xmath49 : @xmath50 given @xmath0 ( the number of bits ) and @xmath1 ( the number of oracle - queries ) , a positive value of @xmath49 indicates that the quantum algorithm is more efficient than the classical one . for @xmath51 and @xmath52
, @xmath49 attains its absolute maximum at @xmath53 and @xmath54 : @xmath55 for fixed @xmath0 , the function @xmath49 has relative maxima at @xmath56 with @xmath57 , for @xmath58 . in fig .
[ fig1 ] , @xmath49 is plotted as a function of @xmath1 for different values of @xmath0 ; the curves displayed from left to right are associated with increasing values of @xmath0 , @xmath59 : as it can be argued from the picture plotted above , @xmath49 is a concave function in the variable @xmath1 for every fixed value of @xmath60 .
therefore , the worst case efficiency of the quantum algorithm increases , with respect to the classical one , for @xmath61 ( where @xmath62 for @xmath58 ) , until it reaches its ( relative ) maximum value .
such a value exhibits a weak dependence on the variable @xmath0 and it is approximately equal to @xmath63 for @xmath58 . on the contrary , for @xmath64
, @xmath49 decreases to zero , thus indicating a substantial equivalence of the two methods for large @xmath1 s .
the above results show that , in the worst case , it is almost always better to resort to the quantum rather than to the classical algorithm .
however , the improvement of quantum queries over the classical ones decreases with the number of queries .
furthermore , @xmath49 becomes negative for @xmath65 , namely there is a ( small ) range of values of @xmath1 for which the classical querying procedure gives less probability of error than the quantum one .
not surprisingly , this is due to the fact that the @xmath66-th classical query is able to determine with certainty if the function is constant or not , since @xmath67 .
instead , the quantum algorithm , as it has been conceived , can never answer with certainty since @xmath47 never vanishes .
until now we have restricted our analysis to the worst possible scenario , i.e. when the function @xmath7 outputs the same value for all possible inputs but one . in the general case , the function may be any of the @xmath4 possible @xmath0-bit boolean functions belonging to the set @xmath8 .
this set contains @xmath31 constant functions ( one always outputs @xmath3 , while the other outputs @xmath2 ) , @xmath68 balanced functions ( they output the same value for exactly one - half of the possible inputs ) and @xmath69 @xmath70-type functions ( they output @xmath3 ( @xmath2 ) on @xmath71 inputs and @xmath2 ( @xmath3 ) on the remaining @xmath72 ones ) .
we repeat the analysis of the previous section and suppose we are given one particular @xmath70-type function .
the probability of obtaining always the same output bit by querying the classical oracle @xmath1 times , and thus of concluding , erroneously , that the function is constant , is equal to @xmath73 + \textrm{prob } [ \underbrace{1\dots 1}_{k}]$ ] .
the latter are the probabilities of obtaining @xmath1 consecutive digits @xmath3 and @xmath2 respectively , and they explicitly read @xmath74 = \prod_{j=0}^{k-1 } \frac{2^{n}-m - j}{2^{n}-j}= \frac{(m-2^{n})_{k}}{(-2^{n})_{k}}\:,\ ] ] @xmath75 = \prod_{j=0}^{k-1 } \frac{m - j}{2^{n}-j}= \frac{(-m)_{k}}{(-2^{n})_{k}}\:.\ ] ] in the above , for simplicity of notation , the pochhammer symbol @xmath76 has been used .
therefore , the classical error probability @xmath77 in declaring constant an @xmath70-type function is : @xmath78 the corresponding probability @xmath79 for the quantum case is much easier to calculate .
in fact , given an @xmath70-type function , the conditional probability of obtaining @xmath1 consecutive measurement results @xmath43 via the iterated deutsch - jozsa procedure is , according to eq .
( [ worst4 ] ) : @xmath80 \equiv q_{m}(k , n ) = \big\vert \frac{1}{2^{n } } \sum_{x=0}^{2^{n}-1 } ( -1)^{f(x)}\big\vert^{2k}= \big ( 1-\frac{m}{2^{n-1 } } \big)^{2k}\:.\ ] ] as in the previous section , we can now define a function @xmath81 which quantifies the relative efficiency of the quantum algorithm with respect to the classical one : @xmath82 @xmath81 does not dramatically change with @xmath0 ; therefore , we fix @xmath83 and present typical behaviours for different values of @xmath72 .
we notice that @xmath81 is symmetric around @xmath84 , and that , when @xmath72 gets close to @xmath85 , the function @xmath70 tends to be balanced and @xmath81 rapidly decreases .
this behaviour indicates that the quantum algorithm can perform much better than the classical one , but only for small numbers of oracle - queries . on the contrary ,
the two procedures become equally efficient for larger values of @xmath1 . for instance , in fig .
[ fig2 ] , when @xmath86 and the function @xmath70 is almost balanced , the two procedures are nearly indistinguishable for @xmath87 . instead , in the general case of @xmath72 significantly different from @xmath85 , the dependence of @xmath81 on @xmath1 is much smoother and presents a maximum and a change of concavity .
further , for @xmath65 , @xmath81 assumes negative values , invisible in the above pictures .
this indicates a better performance of the classical algorithm over the quantum one , as already noticed when dealing with the @xmath10-type function in the previous section .
however this phenomenon is much less pronounced than in the worst case situation .
we now tackle the general case and suppose we are given a randomly chosen boolean function belonging to @xmath8 .
then , the average classical error probability @xmath88 in declaring such a function constant , after @xmath1 consecutive queries with identical outputs , is : @xmath89 + \frac{{2^{n}\choose 2^{n-1}}}{2^{2^{n}}}\ , \frac{2(-2^{n-1})_{k}}{(-2^{n})_{k}}\:.\ ] ] in the previous expression , the factor @xmath90 , with @xmath91 , gives the probability of sorting , within the set @xmath8 , an unbalanced @xmath70-type function which outputs a number of zeroes less than @xmath85 ( the extra factor @xmath31 in front of the sum takes into account the contribution coming from functions which output a number of zeroes which is greater than @xmath85 ) .
analogously , the factor @xmath92 gives the probability of sorting a balanced function . on the contrary
, the corresponding quantum error probability @xmath93 of choosing at random a boolean function within @xmath8 and declaring it constant , due to @xmath1 consecutive measurement results @xmath43 , is : @xmath94 as before , we introduce the average efficiency function @xmath95 : @xmath96 + \ ] ] @xmath97 this function quantifies the ( average ) relative efficiency of the quantum algorithm with respect to the classical one : the larger @xmath95 , the more preferable the quantum procedure with respect to the classical one .
the dependence of @xmath95 on @xmath1 and @xmath0 is shown in fig .
[ fig3 ] below , where its different behaviour with respect to @xmath81 in fig .
[ fig2 ] emerges .
it is worth commenting on the behaviour of @xmath95 .
first , we notice that this function is approximately zero for most of the values of @xmath0 and @xmath1 , indicating a substantial equivalence of the two algorithms ; indeed , in all but a few situations , they are able to solve the problem with negligible probability of error .
second , for fixed @xmath0 , @xmath95 is monotonically decreasing in @xmath1 ; for @xmath54 , the quantum error probability @xmath98 falls rapidly to zero , while the classical one , @xmath99 , approaches unity .
the decrease to zero of @xmath98 is due to the fact that the number of balanced ( or almost balanced ) functions increases considerably for @xmath0 large and the quantum algorithm is able , as we already know , to recognize them with certainty ( or almost certainty ) . on the contrary ,
the number of _ worst _ case functions , i.e. those which output almost always the same value , gets rapidly to zero and the probability of sorting one of them becomes negligible .
this argument does not apply to the classical probability @xmath99 , which instead tends to @xmath2 ; indeed , the classical algorithm fails to give the right answer for balanced ( or almost balanced ) functions which form the overwhelming majority in @xmath100 . finally , unlike the previous cases , the numerical analysis indicates no negative values of @xmath95 ; thus , in the average case , the quantum algorithm is expected to be always better than the classical one .
in this paper we have shown how a trivial iteration of the original deutsch - jozsa algorithm , can be applied to solve the general problem of deciding if a function @xmath7 , sorted at random among the set of @xmath0-bit boolean functions , is constant or not .
we have compared the error probability of the quantum procedure with that of the classical one , the latter consisting of consecutive oracle - queries performed until two different outputs are encountered .
we have analyzed the range of values of @xmath0 and @xmath1 for which the quantum procedure is more efficient than the classical one , both on average and in the worst possible situation . from a numerical analysis ,
one concludes that , by iterating the quantum algorithm , one can always solve the problem , on average , _ better _ than with the classical method , and _ much better _ in those situations where the number of queries @xmath1 is small and @xmath66 large .
moreover , we have seen that , when @xmath7 is a @xmath70-type function with @xmath72 not much greater than @xmath2 ( the worst cases ) , then the classical method is asymptotically ( that is , when @xmath66 is large and @xmath101 ) preferable to the quantum one .
this peculiar effect is due to the fact that the deutsch - jozsa algorithm , unlike the classical querying , can never solve the general problem with certainty since it does not allow for conditioning on previous results .
w. van dam ; _ quantum oracle interrogations : getting all information for almost half the price " _ , in _ proceedings of the 39th annual symposium on foundations of computer science ( focs98 ) _ , ieee , ( 1998 ) . | we study the probability of making an error if , by querying an oracle a fixed number of times , we declare constant a randomly chosen @xmath0-bit boolean function .
we compare the classical and the quantum case , and we determine for how many oracle - queries @xmath1 and for how many bits @xmath0 one querying procedure is more efficient than the other .
addtoresetequationsection keywords : boolean functions ; oracles ; deutsch - josza algorithm . |
the functionalization of colloids allows fine tuning the directionality of the interactions , an important tool for material design.@xcite recently , there has been a significant advance on both the production of these particles@xcite and the theoretical understanding of their equilibrium phase diagrams .
@xcite however , the kinetics of aggregation and self - organization are still largely unexplored .
pioneer studies of aggregation of patchy colloids in the limit of irreversible binding have revealed non - equilibrium structures which significantly differ from the thermodynamic ones.@xcite however , while in equilibrium the obtained structures do not depend on the history of the assembly process , under non - equilibrium conditions this is not the case . the formation of strong bonds between particles , characterized by very long relaxation times , yields huge energy barriers for the relaxation towards equilibrium .
consequently , kinetically trapped structures are obtained which imprint the kinetic pathways of aggregation . here , we compare two limiting cases of mass transport in solution : advection and diffusion . the number of bonds that a patchy colloid can form controls the phase diagram of the assembly .
@xcite mixtures of patchy colloids add novel features to the behavior of these complex fluids , e.g. , the formation of interpenetrating gels or bigels.@xcite in particular , mixtures of two- and three - patch colloids have been considered as they allow control of the ratio between chains and branches , through the fraction of two- and three - patch particles , respectively .
theoretical equilibrium phase diagrams of these mixtures reveal a drastic reduction of the phase separation region and the possibility of equilibrium low density liquids.@xcite the interest in the aggregation on substrates is twofold .
first , from the practical point of view , a substrate works as a nucleation center for growth improving the controllability over assembly.@xcite second , with substrates it is possible to define a growth direction ( away from the substrate ) and characterize the time evolution of the structure .
this possibility constitutes a powerful tool for a systematic theoretical study of non - equilibrium growth @xcite .
we performed extensive kinetic monte carlo simulations of the irreversible aggregation on a substrate in the limit of two mechanisms of mass transport : diffusion and advection .
we show that in the limit of diffusive transport , an optimal ratio of two- and three - patch colloids is found that maximizes the density of the structure .
by contrast , under advection , the density decreases monotonically with the fraction of two - patch colloids .
this article is organized as follows : in sec .
[ sec.model ] we introduce the model and the underlying physical considerations , as well as the simulation details ; in sec .
[ sec.results ] the results for solutions of three - patch colloids ( [ sec.three ] ) and mixtures of two- and three - patch colloids ( [ sec.two-tree ] ) are presented . finally , in sec . [ sec.conclusions ]
we draw some conclusions .
, title="fig : " ] + we use the model proposed in ref
. @xcite .
colloids are considered spherical particles with @xmath0 patches equally spaced on their surface .
interactions are assumed pairwise , repulsive between the spherical cores and attractive between patches . in both cases
the interaction is short ranged , such that the colloid - colloid interaction can be described as excluded volume interaction and only two , almost overlapping , patches can bind .
below , we describe in detail the patch - patch interaction as well as the two limits of mass transport in solution : diffusion and advection ( see fig . [ fig::brownvsball ] ) .
the patch - patch interaction is considered short range .
thus , we assume that patches of two different colloids can only bind upon collision between colloids . for each patch
, we define the interaction range on the surface of the colloid , around the patch , describing the area within which the patch - patch interaction is effective . as shown in fig .
[ fig::interaction.range ] , we define the interaction range ( green ) of the patch ( red ) as a circular region on the surface of the colloid , parametrized by the angle @xmath1 . if the contact point upon collision is within the interaction ranges of both patches , binding is successful , otherwise an elastic collision occurs . in the simulation , as only one
colloid moves at a time , collision occurs between the moving colloid and one colloid bound to the growing structure . in this case
, binding occurs with probability @xmath2 when the new arriving colloid collides within one of the interaction ranges of the other .
@xmath2 is the fraction of the surface of the new colloid covered by the interaction range of one of its patches .
we have chosen @xmath3 .
the interaction range accounts for both the patch extension and the range of the patch - patch interaction .
collisions with the surface always result on the adsorption of the colloids . in this model
, we consider chemical bonds between patches .
this implies that bonds are highly directional and therefore promote the alignment of the colloids along their patches .
additionally , chemical bonds are typically characterized by binding energies larger than the thermal energy .
thus , we assume that binding is irreversible within the timescale of interest . with the center of the patch .
[ fig::interaction.range ] , title="fig : " ] + diffusion in solution occurs through successive collisions with the solvent .
we consider particles to diffuse one at a time towards the substrate .
the collisions with the solvent are described as a poisson process with a time between collisions , @xmath4 , exponentially distributed @xmath5 where @xmath6 is the collision rate and @xmath7 the probability distribution . at each colloid / solvent collision
a new velocity is assigned to the colloid with the direction uniformly distributed and the magnitude @xmath8 following the maxwell - boltzmann distribution at the thermostat temperature @xmath9 , @xmath10 where @xmath9 has units of @xmath11 , @xmath12 is the mass of the colloid and @xmath13 the boltzmann constant .
it is then possible to fine tune the diffusion coefficient @xmath14 by selecting the adequate thermostat temperature and collision rate . to study advective mass transport
, we describe the trajectory of the colloids towards the substrate as a straight vertical line ( see fig .
[ fig::brownvsball](b ) ) .
we randomly generate a position in the horizontal plane and perform a downward vertical movement until the particle either collides with the substrate , sticking to it , or to another particle . as in the diffusive case , binding with another particle occurs when their interaction ranges overlap .
first we discuss the case of three - patch colloids and then proceed with the study of mixtures of two- and three - patch colloids . .
( inset ) liquid - film density as a function of the substrate lateral size.,title="fig : " ] + as a function of the height of the film @xmath15 for diffusive ( open squares ) and advective ( open circles ) transports . measurements were performed for the system of three - patch colloids on a substrate of linear size @xmath16 .
, title="fig : " ] + the aggregation of three - patch colloids on a substrate under diffusive transport has been studied in ref.@xcite . in that work ,
three regimes in the density profile of the colloidal network were identified : the surface layer , dominated by the influence of the substrate ; the liquid film , where the density of the network of colloids saturates at @xmath17 ; and the interfacial region , corresponding to the only active region of the structure .
figure [ fig.ballistic_size_nowall ] shows the density profile for the advective transport of three - patch colloids revealing also three different regimes . as we will discuss below , although these regimes are the same as those obtained with diffusion , their structural properties are significantly different . in the diffusive case ,
the structure of the colloidal network is fractal with fractal dimension @xmath18 .
this value is consistent with the one found for diffusion limited aggregation ( dla ) and diffusion limited deposition ( dld)@xcite .
the density @xmath19 scales as a power law with the lateral size l of the substrate , with the prefactor depending on the diffusion coefficient @xmath14 . in the advective case ,
no significant finite - size dependence is observed , as shown in fig .
[ fig.ballistic_size_nowall ] .
this result indicates that the colloidal network is not fractal in this case . on a substrate of linear size @xmath16 .
snapshots of a region in the liquid film for different fractions of two - patch colloids , @xmath20 . from left to right , top to bottom , @xmath20 is @xmath21 , @xmath22 , @xmath23 , and @xmath24 .
three - patch colloids are in blue ( dark ) , two - patch colloids are in green ( light ) , the ( red ) spheres on the surface of the colloids represent the patches and the ( red ) sticks the connections between them .
[ fig.dimers_snaps_brown ] ] as the colloidal network grows away from the substrate , some branches will grow faster than others and shield the access of new colloids to the inner ones , hindering further binding , and the growth of these branches . the competition between branches , however , depends on the mechanism of mass transport .
figure [ fig.cluster ] shows the number of different branches @xmath25 as a function of the distance @xmath15 to the substrate . in both cases
, the number of branches asymptotically converges towards one and therefore we subtract the values in the vertical axis by unity . in both cases , @xmath25 decays exponentially .
nevertheless , the decay is faster under diffusion .
the erratic motion of diffusing colloids promotes the lateral growth of the network and favors shielding of the inner branches . in the advective case , trajectories are strictly downwards and lateral growth is much slower .
this is also evident from the largest value of z for advective transport , for the same number of deposited particles . on a substrate of linear size @xmath16 .
snapshots of a region in the liquid film for different fractions of two - patch colloids , @xmath20 . from left to right , top to bottom , @xmath20 is @xmath21 , @xmath26 , @xmath23 , and @xmath24 .
three - patch colloids are in blue ( dark ) , two - patch colloids are in green ( light ) , the ( red ) spheres on the surface of the colloids represent the patches and the ( red ) sticks the connections between them .
[ fig.dimers_snaps_ball ] ] ] when mixtures of two- and three - patch colloids are considered , two mechanisms compete : the formation of long chains favored by two - patch colloids and branching promoted by three - patch ones .
we define the variable @xmath20 as the fraction of two - patch colloids . as shown in fig .
[ fig.dimers_snaps_brown ] , the larger the @xmath20 the longer the chains of two - patch colloids .
it was shown in ref.@xcite that , under diffusive transport , the network is always fractal for any @xmath27 .
the measured fractal dimension is consistent with that obtained for three - patch colloids . in fig .
[ fig.dimers_snaps_brown ] , a density maximum is observed at a fraction of two - patch colloids around @xmath28 , by contrast to equilibrium coexisting gels where a monotonic decrease is observed @xcite .
this maximum is related to the competition between the formation of long chains and the maximization of accessible patches.@xcite the influence of advective transport on the growth of a mixture of two- and three - patch colloids is discussed next .
the density decreases monotonically with @xmath20 , as shown in fig .
[ fig.dimers_snaps_ball ] .
this behavior is similar to that observed under equilibrium.@xcite however , in the nonequilibrium case the monotonic decrease is related to the mechanism of mass transport .
colloids only follow vertical trajectories towards the substrate and can not circumvent the branches and squeeze into the fjords as in diffusive transport .
therefore , an increase in the fraction of two - patch colloids does not favor the access to available patches in the bulk and only promotes the growth of linear chains at the forefront of the network . in spite of observing similar qualitative behavior for the density in the equilibrium and advective cases , in the nonequilibrium case the growth direction ( away from the substrate )
is imprinted in the branching . in fig .
[ fig.equilibrium ] we compare the density of the nonequilibrium film with the equilibrium liquid density at coexistence and on the spinodal .
the coexistence curve corresponds to the region of the phase diagram where liquid and vapor coexist .
the spinodal line delimits the region of metastability within the coexistence curve . between the coexistence and the spinodal lines the liquid is mechanically stable .
we note that both equilibrium densities decrease monotonically with @xmath20 .
interestingly , the nonequilibrium densities almost overlap the liquid densities along the spinodal line .
significant differences are observed only in the limit of small ratio of two - patch colloids , where the density at the spinodal is slightly higher .
note that we are comparing the nonequilibrium density on a substrate with equilibrium bulk densities .
however , as shown in ref.@xcite , the properties of the network in the liquid film regime are not sensitive to the interaction with the substrate , in line with the behavior of nonequilibrium wetting films .
@xcite as described earlier , the maximum in the density observed for diffusive transport is likely driven by the increasing accessibility to empty patches with @xmath20 . to check this hypothesis
we perform a detailed analysis of the film growth mechanisms .
the colloidal network grows from several chains starting on the substrate which either branch or stop growing at some point . branching is expected to occur at three - patch colloids . for each of these particles , in the network
, one patch is connected to the chain coming from the substrate .
thus , we measure the correlation between the states of the other two patches .
we define the state of a patch as : empty , if there is no colloid connected to it ; trimer , if it is connected to another three - patch colloid ; or s - chain , if it is connected to a linear chain of @xmath29 two - patch colloids . in the left column of fig .
[ fig.cc_analysis ] we plot the distribution of four different configurations . from top to bottom , we plot the size distribution of the neighboring chain of an empty patch , a trimer , a 1-chain , and a 2-chain .
we found that the distribution of colloids depends on the mechanism of mass transport . in diffusive transport , for configurations with an empty patch of reference ( first line of fig .
[ fig.cc_analysis ] ) , there is a higher probability for the neighboring patch to be also empty .
this is related to the caging of colloids , as some colloids are trapped in cages surrounded by other chains , which hinder the access of new colloids to any of the patches .
increasing the ratio of two - patch colloids , increases the size of such cages and , thereby , the probability that a diffusing colloid will end up there . for the other configurations , the probability of the patch being empty is lower , since , if the colloid is not inside a cage , the diffusive transport allows many paths for colloids to bind to both patches . in advective transport , a qualitatively different picture is observed .
if the reference patch is empty , there is a lower probability for the neighboring patch to be also empty , when compared with the diffusive case . this probability does not change with the ratio of two - patch colloids since , regardless of the available volume around the empty patch , a patch that is screened by a chain on top of it will never be reachable during a downward trajectory of the particles . for the other configurations ,
the neighbor patch is mostly empty , as a result of the orientation of the patch and the screening effect .
this is supported by the distribution of chains for @xmath30 , where for diffusive transport , the larger the chain , the lower the probability of finding an empty patch . by contrast , for advective transport , the probability of finding an empty patch remains constant for @xmath31 and @xmath32 .
] finally , fig . [ fig.cca_average ] depicts the average size @xmath29 of the chains of two - patch colloids .
the chains for diffusive transport are typically larger than those for the advective case .
note that , for advective transport most of the neighboring patches of an occupied or empty patch are also empty , significantly decreasing the average size of the chain ( see fig .
[ fig.cc_analysis ] ) .
the reason is that branching implies that both available patches of a three - patch colloid are aligned with downward trajectories .
this is rarely the case , as the fluctuations in the orientation of the chain and of the colloid at the tip will typically shield the patches , hindering branching .
the picture is significantly different under diffusion . diffusing particles can circumvent colloids and connect with shielded patches , favoring branching and lateral growth .
we have shown that the kinetics of mass transport towards the substrate significantly affects the structure of the network of patchy particles .
this is in deep contrast with growth under equilibrium conditions where the history of growth is irrelevant .
we observe that , diffusive transport favors branching , lateral growth , and the formation of a fractal network of colloids , yielding an optimal fraction of two - patch colloids at which the film density is maximized . for advective transport ,
the network is not fractal and the density decreases monotonically with the fraction of two - patch colloids .
interestingly , we found that the liquid film density , for irreversible adsorption with advective transport , is very similar to the one previously observed for the spinodal curve at equilibrium , albeit their structures are quite different .
the detailed reason for such similarity is an open question , although it may be related to the mechanical instability of the liquid at the spinodal .
our study highlights the dependence of the liquid film structure on the growth conditions .
however , we considered the limiting case where binding is irreversible . for significantly longer time scales and/or high temperatures ,
the possibility of bond breaking should be considered and relaxation towards equilibrium is to be expected .
future works should consider the dynamics of such relaxation .
46ifxundefined [ 1 ] ifx#1 ifnum [ 1 ] # 1firstoftwo secondoftwo ifx [ 1 ] # 1firstoftwo secondoftwo `` `` # 1''''@noop [ 0]secondoftwosanitize@url [ 0 ] + 12$12 & 12#1212_12%12@startlink[1]@endlink[0]@bib@innerbibempty @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop
_ _ ( , , ) @noop * * , ( ) @noop * * , ( ) in http://adsabs.harvard.edu/abs/2004cond.mat..9239s[__ ] ( , ) pp . | patchy particles are a class of colloids with functionalized surfaces . through surface functionalization , the strength and directionality of the colloidal interactions are tunable allowing control over coordination of the particle .
exquisite equilibrium phase diagrams of mixtures of coordination two and three have been reported . however , the kinetics of self - organization and the feasibility of the predicted structures are still largely unexplored . here , we study the irreversible aggregation of these mixtures on a substrate , for different fractions of two - patch particles . two mechanisms of mass transport are compared : diffusion and advection . in the diffusive case , an optimal fraction is found that maximizes the density of the aggregate .
by contrast , for advective transport , the density decreases monotonically with the fraction of two - patch colloids , in line with the behavior of the liquid density on the spinodal of the equilibrium phase diagram . |
the chemical evolution of the solar neighborhood , as a prototype of galactic disks in general , has been an active field of astrophysical research over the last four decades , starting from the pioneering works of van den bergh ( 1962 ) , schmidt ( 1963 ) , and pagel & patchett ( 1975 ) . with talbot & arnett ( 1971 ) and then tinsley ( 1980 ) , the subject got its elegant physico - mathematical formulation , then adopted in most later works . in this traditional approach chemical evolution is rigorously treated from first principles , adopting specific initial and boundary conditions , along with heavy element yields obtained from theoretical stellar model and supernova explosion calculations . in this paper
i instead adopt a purely phenomenological approach , lining up a series of observational facts and then building on them semi - quantitative inferences .
this should be regarded as complementary to the traditional approach , and may perhaps drive it towards unexplored regions of the theoretical parameter space .
i also adopt a ` simple minded ' attitude , assuming when appropriate that the most straightforward interpretation of the facts is the right one , and no _ cosmic conspiracies _ are at work ( for example , it is subintended that the imf is universal )
. for conciseness the presentation will be somewhat schematic .
clusters are perhaps the best realization in nature of the closed - box model of chemical evolution . within clusters
we find confined in the same place all the dark matter , all the baryons , all the galaxies , all the stars and all the metals , that have participated in the play .
clusters are then good _ archives _ of their past star formation ( sf ) and metal production history .
x - ray observations of clusters provide iron and @xmath4-element abundances in the icm , along with the icm mass .
optical observations provide the luminosity , mass , and the average metallicity of all the stars now locked inside galaxies ( and even hints to a trace population of intergalactic , free - floating stars ) .
optical and near - ir observations of cluster elliptical galaxies have also set tight limits to the formation epoch of the bulk of the stars contained in them .
cumulatively , these observations have then established the following facts ( see renzini 1997 , 1999a , b and extensive references therein for the original sources ) : the average iron abundance in the icm is @xmath5 .
iron mass to light ratio _ is @xmath6 , where @xmath7 , @xmath8 is the mass of the icm , and @xmath9 the total @xmath10-band luminosity of all cluster galaxies . the average iron abundance of cluster stars is roughly solar .
the total iron mass to light ratio is @xmath11 , only weakly dependent on the hubble constant .
the global elemental ratio @xmath12fe ] in the icm+stars is roughly solar , hence the total cluster _ metal mass to light ratio _ is @xmath13 , since in solar proportions @xmath14 .
this is a fully _ empirical _ estimate of the metal yield of a now old stellar population .
most metals are out of galaxies : @xmath15 , i.e. there is @xmath16 times more iron in the icm than there is iron locked into stars , for @xmath17 , or @xmath18 times more for @xmath19 .
massive starbursts promote major , metal - enriched galactic winds ( heckman et al .
most stars in galaxy clusters belong to galactic spheroids ( ellipticals and bulges ) , and formed in massive starbursts at @xmath20 .
note that in @xmath21 both quantities are measured _ now _ ; however , the metal mass @xmath22 was produced and released at very early times by the stellar population that after aging for @xmath23 gyr has faded to the luminosity @xmath9 . in the adopted ` simple - minded '
approach , from these empirical facts the following inferences can be drawn : @xmath24 most metals in clusters ( now partly in the icm , partly in stars ) were produced at @xmath20 .
@xmath24 icm metals were ejected by starburst driven galactic winds at the production time , i.e. at @xmath20 .
to which extent are clusters _ fair samples _ of the universe as a whole ?
they are certainly the highest density peaks in the distribution of matter , both dark and shining , and one may expect sf and chemical evolution to have proceed differently in clusters compared to the low - density _
field_. however , there are also striking similarities , including the following ones : most stars are now in galactic spheroids , in the field as in the clusters ( e.g. up to @xmath25 according to fukugita , hogan , & peebles 1998 ) .
the fraction of baryons now locked into stars is nearly independent of the environment , being @xmath26 in both clusters and field ( e.g. renzini 1997 ; fukugita et al .
the sf histories may well have been different , confined at early times in clusters , more protracted in the field , but the @xmath27 endproducts appear to be very similar , with nearly the same efficiency of baryon to star conversion . field ellipticals and s0 s are very similar to cluster ellipticals and s0 s , with the luminosity - weighed age of their stellar populations being at most @xmath28 gyr less than that of the cluster galaxies ( bernardi et al .
the stellar populations of ( large ) bulges are very similar to ellipticals .
the majority of them appear to follow the same mg@xmath29 relation of ellipticals , while a minority of outliers likely had a significant episode of star sf at later times ( jablonka , martin , & arimoto 1996 ) .
the stellar population of the galactic bulge is nearly as old as galactic globular clusters in halo , or 13 - 15 gyr ( ortolani et al . 1995 ) .
again , from this second series of facts more ` simple - minded ' inferences can be drawn : @xmath24 having the sf proceeded to the same @xmath26 baryon to stars conversion in the general field and in clusters , the global metallicity of the @xmath27 universe has to be nearly the same as that we can actually see in clusters , i.e. @xmath30 solar . @xmath24
the bulk of stars in galactic spheroids in clusters as well as in the field formed at @xmath20 . @xmath24 since some 50 to 75% of all stars are now in spheroid , and the bulk of stars in spheroids formed at @xmath20 , the inference is that at least @xmath30 of all stars formed at @xmath20 . this ` fossil evidence ' argument agrees with the direct determination of the sf history at high redshift ( steidel et al . 1999 ) .
@xmath24 therefore , the metallicity of the @xmath31 universe was @xmath30 of its present value , i.e. @xmath32 solar , a _ prompt initial enrichment of the universe _
( renzini 1999a ) .
when did the morphological differentiation of galaxies take place ?
when did the disks of the present - day spirals started to be assembled ? here
are some facts : we see no disk galaxies at @xmath20 .
lyman break galaxies appear to be much smaller , compact objects , most likely the progenitors of today s objects that the fossil evidence date having formed at @xmath20 , i.e. galactic spheroids ( e.g. giavalisco , steidel , & macchetto 1996 ) .
spirals are well in place by @xmath33 , along with passively evolving ellipticals ( e.g. abraham & van den bergh 2001 ) . hence :
@xmath24 bulges come first , formed in starbursts , and disks are slowly added later , if the environment is quiet enough to allow for their formation and survival . @xmath24 the morphological differentiation of galaxies ( the emergence of the hubble sequence ) took place between @xmath34 and @xmath35 , i.e. in the so far poorly explored _ desert _
range @xmath36 .
in the k band the galactic bulge and disk contribute respectively @xmath37 and @xmath38 ( kent , dame , & fazio 1991 ) , and in the @xmath10 band the bulge luminosity is @xmath39 .
> from the cluster empirical yield it follows that the galactic bulge has produced @xmath40 of metals . where are all these metals
? two billion solar masses of metals should not be easy to hide , yet ... the stellar mass of the bulge follows from its @xmath41-band mass to light ratio , @xmath42 ( kent 1992 ) , and its luminosity , hence @xmath43 .
its average metallicity is about solar(mcwilliam & rich 1994 ) , i.e. @xmath44 , and therefore the bulge stars all together contain @xmath45 of metals .
only @xmath1 of the metals produced when the bulge was actively star forming some 13 gyr ago are still in the bulge !
this implies that @xmath46 , or @xmath47 were ejected into the surrounding space by an early wind .
a word of caution is in order .
the estimated the metal yield ( section 2 ) follows from adopting @xmath48 solar .
more recent estimates prefer @xmath49 solar ( de grandi & molendi 2001 ) , the total metal production by the bulge reduces to @xmath50 , of which @xmath0 had to be ejected .
the bottom line is that at least @xmath51 times more metals were ejected , than retained in the bulge . at the time of bulge formation ,
such @xmath0 of metals run into largely pristine ( @xmath52 ) material , experienced r - t instabilities leading to chaotic mixing , and establishing a distribution of metallicities in a largely inhomogeneous igm surrounding the young bulge .
for example , this enormous amount of metals is able to bring to a metallicity 1/10 solar ( i.e. @xmath53 ) about @xmath2 of pristine material , several times the mass of the yet to be formed galactic disk .
if we look to the formation of the milky way galaxy from a purely empirical point of view , we have an old , now passively evolving bulge , and a younger disk , still forming stars even if likely at a reduced rate compared to a more active past .
hence , we can distinguish three main phases in the formation and evolution of our own galaxy .
, some 13 gyr ago ( `` at @xmath34 '' ) .
the relatively fast ( @xmath54 gyr ) assembly of the bulge from smaller subunits promotes a massive starburst ( sfr=10 - 100 @xmath55yr ) , which drives a metal rich wind and @xmath0 of metals are ejected into the surrounding medium .
, lasting a poorly constrained lapse of time ( some gyr ?
, until @xmath56 ? ) , during which the bulge settles into its passive evolution , the ejecta partially mix with the surrounding medium , and a mass some 5 - 10 times larger than the present mass of the galaxy is inhomogeneously contaminated , with its average metallicity being raised to @xmath1 solar .
heating of this metal enriched environment ( mee ) by the early bulge wind , may prevent from a while further growth of the stellar mass of the protogalaxy .
, from @xmath56 to @xmath35 ( ? )
. around the aging bulge , the mee is cooling , infall starts of @xmath57 solar material from the mee , and the galactic disk begins to form and grow .
these semi - quantitative arguments may help settling on a final solution for an old problem . from the earliest times mentioned in the introduction , the existence of a `` g - dwarf problem '' was soon recognized : the solar neighborhood contains far too few metal poor stars ( @xmath58 solar ) compared to the prediction of a simple , closed box model of chemical evolution ( see pagel 2001 , for a recent discussion of the problem ) .
traditional solutions of the g - dwarf problem have considered three options ( or some combination of them ) : an initial _ top heavy _
imf , a _ prompt initial enrichment _ ( pie ) model , and infall , i.e. the gradual assembly of the disk as opposed to a zero metallicity disk fully assembled from the beginning .
the above considerations clearly indicate that pie is inevitable , once one accepts that the bulge comes first and the disk later .
note that infall models have often assumed @xmath52 for the infalling material .
some gradual growth of the disk is also inevitable , being the alternative assumption of instantly assembled disk quite unrealistic .
so , a combination of pie and infall emerges as the natural solution of the old problem .
worth mentioning here is that the idea of the bulge pre - enriching the disk is certainly not new ( see e.g. kppen & arimoto 1990 ) .
this sketchy cartoon of galaxy formation is hard to put in more quantitative terms , purely from first principles .
models would have difficulties to predict the extension of the mee , its distribution of metallicities , the duration of the intermission , the evolution of the infall rate and the disk build up .
all these phenomena involve highly non - linear , weather - like hydrodynamics with little predictive power .
however , we are living in a special time for astronomy , and where theory gets stuck observations can help .
we can actually start to _ see _ directly what theory is not able to predict . with the current generation of large ground based telescopes , together with the x - ray , optical , and infrared facilities now or soon in space , we have for the first time the possibility to empirically map the growth of galaxies , to follow back in time the disappearance of the hubble sequence , and its emergence , forwards from high to low redshifts . | scaling from the empirical metal yield as measured in clusters of galaxies , it is inferred that early in the evolution of the galaxy the bulge stellar population has produced @xmath0 of metals , at least 5 times more than the total metal content of the bulge today .
it is argued that an early galactic wind from the starbursting bulge has pre - enriched a vast region around it , with these metals being able to enrich to @xmath1 solar of order of @xmath2 of pristine material . from the empirical evidence that bulges come before disks
, it is inferred that the milky way disk formed out of this pre - enriched material , which accounts for the scarcity of metal poor stars in the solar neighborhood , the so - called ` g - dwarf problem ' .
high redshift observations are now becoming able to efficiently explore the @xmath3 region of the universe , when disk formation and morphological differentiation may have taken place .
# 1_#1 _ # 1_#1 _ = # 1 = 11 versim#1#22.9truept = 12 1.25 in .125 in .25 in |
cometary nuclei display a wide variety of terrain and geological features , ranging from pits and mountains to flat or terraced regions .
the rolis descent images of 67p / c - g show grains of regolith ranging from centimeter sizes to several meters , @xcite . behind these meter - sized obstacles wind tails
are resolved in the rolis images .
smaller sized particles as seen by rolis have been observed in orbit by the osiris instrument , @xcite , suggesting that dust particles contribute to both , the cometary tail originating from the nucleus and to the shaping of surface structures .
it has been proposed that the wind - tail structures are the result of boulders blocking impinging particles , which lead otherwise to abrasion of the surrounding terrain .
no model or explanation for the impinging - particle direction indicated by the wind tails has been given , besides the hypothesis that a particle source exists at some distance ( @xcite ) . here
, we establish a predictive model for the directions of impinging particles , which does away with a single particle - source and follows the opposite hypothesis : the whole cometary surface acts as source for dust grains .
the particular shape of the nucleus in combination with the coriolis force leads then to prevailing transport directions in specific regions of the comet .
the same homogeneous activity model has been applied to dust particles lifted off the mantle with higher velocity and yields excellent agreement with the collimated dust structures seen by rosetta around 67/c - g within a few cometary radii , @xcite .
the homogeneous surface - activity model works with a minimal set of parameters and assumptions , summarized as follows : ( i ) incorporation of cometary topography on a scale down to @xmath0 m to ensure a dense and uniform sampling of the entire terrain , ( ii ) accurate computation of the gravitational potential for the partially concave shape of the nucleus , and ( iii ) full inclusion of the rotation of the nucleus .
in particular the rotation of the nucleus has a decisive role for establishing dust transport - vectors on the surface and for higher velocities into space .
slower dust trajectories ( @xmath1 m / s ) are strongly affected by the velocity dependent coriolis force , which pushes particles on circular orbits within a few kilometers around the comet .
in a coordinate - system attached to the cometary body , the acceleration of a dust particle is approximated by @xmath2 which includes the acceleration of dust embedded in the more rapidly expanding gas , the gravitational force and the effect of rotation ( centrifugal and coriolis forces ) .
additional forces , such as radiation pressure , changes in the rotation period or orientation of the rotation axis are neglected .
for instance , the acceleration due to radiation pressure on a 0.1 mm sized spherical dust particle at the perihelion ( 1.25 au ) is one order of magnitude smaller than the gravitational attraction by the nucleus .
we compute the gravitational potential of the irregular and partly concave shape of 67p / c - g by using a @xmath3 triangles polyhedral representation .
assumptions are a homogeneous density of the nucleus and a total mass of @xmath4 kg , @xcite .
the polyhedral gravity equations given by @xcite are numerically efficiently implemented and the dust trajectories are integrated on parallel computing devices using a fourth order runge - kutta integration scheme .
in contrast to previous simulations done by @xcite we fully include the rotation of the comet .
in addition , the integration includes the gas - dust interaction , which has been neglected by @xcite . a stationary gas model around 67/c - g
is constructed by assuming homogeneous outgassing activity across the whole cometary surface , @xcite , and thermal gas velocity ( @xcite , eq .
3.33 ) , @xmath5 molecules at @xmath6 k ( @xcite ) this yields @xmath7 m / s . the co@xmath8 release by 67p / c - g has been detected by the rosina instrument on rosetta at a heliocentric distance of 3.3 au , @xcite . besides the velocity , the second parameter entering the haser model is the surface number - density of the gas @xmath9 .
we neglect different outgassing rates that might originate from different illumination conditions on each surface element .
the transport mechanism described below relies on a slightly higher gravitational pull to the nucleus compared to the gas drag encountered by a dust particle around the comet .
the angular velocity vector and period are from @xcite .
gas density and velocity direction across a slice at @xmath10 km through the cometary nucleus ( @xmath11 m@xmath12 ) .
the highest gas density is reached in the concave neck region due to gas contributing from both side of the valley .
, scaledwidth=40.0% ] the resulting gas number density distribution and velocity vectors are shown in fig .
[ fig : gasfield ] .
the maximal gas density is reached within the neck area of 67p / c - g due to the contribution of gas from the two sides of the neck valley , leading to an increased upward dust flow .
the gas drag depends on the dust - particle mass @xmath13 and radius @xmath14 , as well as on the momentum of the impinging gas molecules . for the standard value @xmath15 from @xcite
we obtain an overall factor @xmath16 for the gas - dust interaction @xmath17 where @xmath18 denote the particle radius and mass . for re - colliding particles ,
the gas - drag forces pointing away from the comet are eventually compensated by the gravitational pull of the comet .
a re - collision is only possible if dust particles are emitted with a non - zero initial velocity acquired in the porous mantle layer of the comet , where gas - drag effects are amplified due to the compressed volume available for expansion , @xcite .
we study dust transport occurring over an illuminated cometary surface with homogeneous gas emission .
day / night effects of changing gas - density due to diurnal varying gas pressure are not considered .
this limits the model to trajectories lasting less then half a cometary rotation period of about 6 h. for the derived global dust transport map , we consider the complete comet as uniformly illuminated and dust emitting .
hence , each trajectory represents a possible dust transport path over the sunlit part of the surface .
the irregular shape of the comet results in collisions with elevated terrain rotating into the pathway .
in addition , trajectories are affected by the modulations in the gravitational potential .
the cometary surface is represented by a mesh of 39996 equal - sized triangles and each center of a triangle serves as origin for dust particles . for the simulation
, we chose an initial velocity as @xmath19 m / s along the triangle outwards normal to sample particles emitted with less than the escape velocity .
the results are independent of the number of triangles used to represent the comet shape and have been reproduced with the 19806 triangles mesh discussed in @xcite .
the maximum gas particle density near the surface is varied from @xmath20 m@xmath12 to @xmath21 m@xmath12 for particles of density @xmath22 kg / m@xmath23 and radius @xmath19 mm , resulting in an overall gas drag very close to the gravitational force .
this procedure selects particles hovering up to several hours over the surface .
note that only the overall prefactor @xmath24 is of importance for the gas drag , and thus , equivalently different particle sizes and gas densities are covered in the parametric study .
the results shown here for co@xmath8 hold also for a h@xmath8o gas atmosphere at @xmath25 k with the @xmath26 ratio decreased by the factor @xmath27 , see eq . ( [ eq : pfgas ] ) . for a coarse - grained and ensemble - averaged picture , we embed the cometary shape in a volume mesh of cubes with side length @xmath28 m to track particles emitted from several surface triangles . in each cubic subvolume
, we establish the average particle momentum direction and assign a total particle momentum of all traversing particles .
in addition , the number of particles and the velocities ( direction and magnitude ) are recorded .
the gas - drag forces delay the re - collision of the particle and lead to sustained motion above the surface while the comet rotates .
the gas forces are expected to temporally vary depending on the night / day exposure of the surface . to focus on trajectories across always sunlit terrain , we eliminate trajectories lasting longer than 6 h ( half the rotation period of the comet )
. we account for the expected downfall of particles caused by a reduction in up - ward gas drag , for instance by locally varying outgassing rates , self - shadowing , varying topographical features , and the day - night terminator by taking into account all trajectories up to a ceiling of 300 m above the surface for the surface impinging particle flux . near surface vectors of dust transport around 67p / c - g .
arrows represent the momentum of the dust transport within the cubic subvolumes closest to the surface .
color ( blue to red ) and arrow size denote increasing momentum per cube .
parameters : maximum gas particle density @xmath11 m@xmath12 , initial dust velocity @xmath29 m / s , homogeneous activity profile across the whole surface . ,
title="fig:",scaledwidth=35.0% ] near surface vectors of dust transport around 67p / c - g .
arrows represent the momentum of the dust transport within the cubic subvolumes closest to the surface .
color ( blue to red ) and arrow size denote increasing momentum per cube .
parameters : maximum gas particle density @xmath11 m@xmath12 , initial dust velocity @xmath29 m / s , homogeneous activity profile across the whole surface .
, title="fig:",scaledwidth=35.0% ] near surface vectors of dust transport around 67p / c - g .
arrows represent the momentum of the dust transport within the cubic subvolumes closest to the surface .
color ( blue to red ) and arrow size denote increasing momentum per cube .
parameters : maximum gas particle density @xmath11 m@xmath12 , initial dust velocity @xmath29 m / s , homogeneous activity profile across the whole surface . ,
title="fig:",scaledwidth=35.0% ] near surface vectors of dust transport around 67p / c - g .
arrows represent the momentum of the dust transport within the cubic subvolumes closest to the surface .
color ( blue to red ) and arrow size denote increasing momentum per cube .
parameters : maximum gas particle density @xmath11 m@xmath12 , initial dust velocity @xmath29 m / s , homogeneous activity profile across the whole surface .
, title="fig:",scaledwidth=35.0% ]
the near surface vectors for the dust transport are shown in fig .
[ fig : velfield ] , where arrows represent the dust transport ( product of velocity and mass ) within the mesh cells closest to the surface .
the largest dust - transfer is predicted in the neck region , despite the homogeneous activity .
the reason is the increased outward gas drag above the concave neck due to contributory gas flow from both sides of the valley .
this process results in an increased upward acceleration and transfer of dust , despite the uniform surface distribution of initial dust grains .
the same effect is predicted in @xcite and observed for the dust jets forming above the neck region .
lll ( a)&(b)&(c ) + prevailing dust transport vectors over the small lobe of 67p / c - g for two different ratios of gas drag / gravity .
panels ( a ) , ( d ) histogram of trajectory lengths . panels ( b ) , ( e ) sample of trajectories reaching a target sphere around the area imaged by rosetta rolis , marked by the red sphere in ( c ) , ( f ) . panels ( c ) , ( f ) volume cell averaged transport vectors above the surface .
( a)-(c ) : gas drag for @xmath30 m@xmath12 supports short - range transport ( mean trajectory length @xmath31 km to the target region ) , while with increased gas drag ( d)-(f ) , @xmath32 m@xmath12 , particles hover over the surface and bridge larger distances ( mean trajectory length @xmath33 km to the target region ) .
the dust trajectories marked by the red arrow in ( e ) coincide with the observed wind tail direction ( white arrow ) .
, title="fig:",scaledwidth=26.0% ] & prevailing dust transport vectors over the small lobe of 67p / c - g for two different ratios of gas drag / gravity .
panels ( a ) , ( d ) histogram of trajectory lengths . panels ( b ) , ( e ) sample of trajectories reaching a target sphere around the area imaged by rosetta rolis , marked by the red sphere in ( c ) , ( f ) . panels ( c ) , ( f ) volume cell averaged transport vectors above the surface .
( a)-(c ) : gas drag for @xmath30 m@xmath12 supports short - range transport ( mean trajectory length @xmath31 km to the target region ) , while with increased gas drag ( d)-(f ) , @xmath32 m@xmath12 , particles hover over the surface and bridge larger distances ( mean trajectory length @xmath33 km to the target region ) .
the dust trajectories marked by the red arrow in ( e ) coincide with the observed wind tail direction ( white arrow ) .
, title="fig:",scaledwidth=25.0% ] & prevailing dust transport vectors over the small lobe of 67p / c - g for two different ratios of gas drag / gravity .
panels ( a ) , ( d ) histogram of trajectory lengths . panels ( b ) ,
( e ) sample of trajectories reaching a target sphere around the area imaged by rosetta rolis , marked by the red sphere in ( c ) , ( f ) .
panels ( c ) , ( f ) volume cell averaged transport vectors above the surface .
( a)-(c ) : gas drag for @xmath30 m@xmath12 supports short - range transport ( mean trajectory length @xmath31 km to the target region ) , while with increased gas drag ( d)-(f ) , @xmath32 m@xmath12 , particles hover over the surface and bridge larger distances ( mean trajectory length @xmath33 km to the target region ) .
the dust trajectories marked by the red arrow in ( e ) coincide with the observed wind tail direction ( white arrow ) .
, title="fig:",scaledwidth=25.0% ] + ( d)&(e)&(f ) + + prevailing dust transport vectors over the small lobe of 67p / c - g for two different ratios of gas drag / gravity .
panels ( a ) , ( d ) histogram of trajectory lengths . panels ( b ) ,
( e ) sample of trajectories reaching a target sphere around the area imaged by rosetta rolis , marked by the red sphere in ( c ) , ( f ) .
panels ( c ) , ( f ) volume cell averaged transport vectors above the surface . ( a)-(c ) : gas drag for @xmath30 m@xmath12 supports short - range transport ( mean trajectory length @xmath31 km to the target region ) , while with increased gas drag ( d)-(f ) , @xmath32 m@xmath12 , particles hover over the surface and bridge larger distances ( mean trajectory length @xmath33 km to the target region ) .
the dust trajectories marked by the red arrow in ( e ) coincide with the observed wind tail direction ( white arrow ) .
, title="fig:",scaledwidth=26.0% ] & prevailing dust transport vectors over the small lobe of 67p / c - g for two different ratios of gas drag / gravity .
panels ( a ) , ( d ) histogram of trajectory lengths . panels ( b ) , ( e ) sample of trajectories reaching a target sphere around the area imaged by rosetta rolis , marked by the red sphere in ( c ) , ( f ) . panels ( c ) , ( f ) volume cell averaged transport vectors above the surface .
( a)-(c ) : gas drag for @xmath30 m@xmath12 supports short - range transport ( mean trajectory length @xmath31 km to the target region ) , while with increased gas drag ( d)-(f ) , @xmath32 m@xmath12 , particles hover over the surface and bridge larger distances ( mean trajectory length @xmath33 km to the target region ) .
the dust trajectories marked by the red arrow in ( e ) coincide with the observed wind tail direction ( white arrow ) .
, title="fig:",scaledwidth=25.0% ] & prevailing dust transport vectors over the small lobe of 67p / c - g for two different ratios of gas drag / gravity .
panels ( a ) , ( d ) histogram of trajectory lengths . panels ( b ) ,
( e ) sample of trajectories reaching a target sphere around the area imaged by rosetta rolis , marked by the red sphere in ( c ) , ( f ) .
panels ( c ) , ( f ) volume cell averaged transport vectors above the surface .
( a)-(c ) : gas drag for @xmath30 m@xmath12 supports short - range transport ( mean trajectory length @xmath31 km to the target region ) , while with increased gas drag ( d)-(f ) , @xmath32 m@xmath12 , particles hover over the surface and bridge larger distances ( mean trajectory length @xmath33 km to the target region ) .
the dust trajectories marked by the red arrow in ( e ) coincide with the observed wind tail direction ( white arrow ) . , title="fig:",scaledwidth=25.0% ] next , we investigate the dust transport vectors around the area imaged during the philae / rolis descent , marked by the red target spheres in fig .
[ fig : velfieldrolis ] ( c ) , ( f ) .
the viewpoint is chosen to approximately match the rolis image area , @xcite suppl .
we select all trajectories entering the target region to identify the origins of impinging dust particles .
typical trajectories are show in fig .
[ fig : velfieldrolis ] ( b ) , ( e ) .
the dust trajectories start perpendicular to the local surface , but then move across the surface and possible around the lobe due to the cometary rotation .
depending on how closely gas drag and gravitational force cancel each other , different travel distances and directions of the grains are predicted .
the distribution of trajectory lengths reaching the target area is given in fig .
[ fig : velfieldrolis ] ( a ) , ( d ) . the trajectories in fig .
[ fig : velfieldrolis ] ( e ) originate at a distance of on average @xmath33 km from the target area along two slightly inclined main transport vectors .
the dust stream indicated by the red arrow in fig .
[ fig : velfieldrolis ] ( e ) provides a possible source of impinging particles to explain the direction of the observed wind tail patterns by @xcite ( white arrow ) .
we have proposed a mechanism for predicting dust migration on the surface of 67p / c - g based on homogeneous dust emission across the entire surface , linked to a detailed topographical model of the nucleus .
the homogeneous model yields a minimal - assumption prediction for the temporally - averaged dust transport .
key aspects are the incorporation of the detailed topography , gravitational potential , and rotational forces of the comet , leading to global `` dust streams '' around 67p / c - g .
previous approaches neglect the rotation of the nucleus due to the small ratio of centrifugal force to gravitational force and gas drag . for dust transfer
, the velocity dependent coriolis effect dominates and rotation of the nucleus can not be neglected .
the predicted transport vectors are in line with the particle directions inferred from the rolis images .
+ the largest dust transport is expected away from the neck region , caused by an increased outward gas drag due to the focused transfer of gas momentum to dust by the concave valley shape .
this mechanism could act in addition to the increased thermal - stress hypothesis proposed by @xcite in the neck region .
+ the same homogeneous activity - model has been applied to the near - nucleus coma by @xcite and is in excellent agreement with rosetta navcam observations of collimated dust jet / structures .
the present approach provides a global and unified theory of dust transfer on the surface and in the coma , in agreement with rosetta observations .
additionally , temporarily / spatially isolated sources of activity can be included in the model , but would require additional parameters acquired by observations .
future observations by rosetta are necessary to map out the global dust transport on 67p / c - g and compare it with predictions from theoretical models .
we thank m. malmer for providing the initial shape model of comet 67p / c - g and the esa / rosetta navcam team for releasing images of 67p / c - g on a continuous basis .
tk acknowledges support by a heisenberg fellowship of the dfg grant kr 2889/5 .
computational resources and support by the north - german supercomputing alliance ( hlrn ) are gratefully acknowledged . | dust transport and deposition behind larger boulders on the comet 67p / churyumov - gerasimenko ( 67p / c - g ) have been observed by the rosetta mission .
we present a mechanism for dust transport vectors based on a homogenous surface activity model incorporating in detail the topography of 67p / c - g .
the combination of gravitation , gas drag , and coriolis force leads to specific dust transfer pathways , which for higher dust velocities fuel the near nucleus coma . by distributing dust sources homogeneously across the whole cometary surface , we derive a global dust - transport map of 67p / c - g .
the transport vectors are in agreement with the reported wind - tail directions in the philae descent area . |
liquid crystals ( lcs ) possess anisotropic interactions because of their molecular shape .
for example , molecules with a rod - like rigid core tend to align parallel to each other and form a mesophase called nematic ; molecules with a disk - like rigid core form discotic phases . in both cases
the rigidity is generated by different combinations of aromatic rings @xcite .
macroscopically , this anisotropy leads to a series of phase transitions that break rotational and translational symmetries in a step - wise fashion .
because of their capacity to reorient , also in response to external fields , lcs are used in a wide range of applications : from the ubiquitous electronic displays , to microlasers @xcite and lubricants @xcite ; but they are also rising to an important role in biomedical sciences and applications @xcite , and in our comprehension of morphogenesis and evolution of living organisms @xcite .
hydrodynamic flow can also couple with the local preferential direction ( director ) established in nematic lcs @xcite .
recently , sengupta _
et al_. have extended microfluidic applications to anisotropic fluids and have found surprising topologies of the orientational field @xcite , and explored fluid and colloidal transport @xcite .
these effects originate from the intimate connection between lc rheological properties @xcite and their local alignment , which can be easily controlled .
however , the complex interplay of confinement to a mesoscopic scale , i.e. of the order of @xmath0 m , surface interactions , hydrodynamic flow and generation of topological defects still poses formidable challenges to both theoretical and experimental investigations .
the theoretical description of lcs based on static continuum theory started in the early 1920s with the work of oseen @xcite , zocher @xcite , and frank @xcite .
the earliest dynamics theory of lcs can be dated back to 1931 by anzelius @xcite . in the 1960s
ericksen and leslie developed a hydrodynamics theory @xcite based on the lc s velocity field @xmath1 and a unit vector describing the local director @xmath2 .
the nematodynamic equations of the ericksen - leslie model are widely used but rest on the assumption that the nematic order parameter is a constant and the nematic lc is uniaxial , and therefore they can not describe physical situations where there is a strong variation of the the nematic order , such as the isotropic - nematic phase transition and the dynamic of topological defects . for cases where there is a strong variation of the nematic order
a tensorial description is necessary , such as the beris
edwards formulation @xcite or the qian
sheng formulation @xcite .
these two approaches differ in the form of the elastic free energy , that is , the former considers the elastic free energy in the one - constant approximation , while the latter allows for two different elastic constants .
we are not aware of any tensorial description of nematodynamics in terms of all three elastic constants , _
i.e. _ splay , twist and bend .
microscopic models have in general the advantage of providing detailed dynamics at small spatial and temporal scales .
molecular dynamics simulations are well suited for this task
. however , physical phenomena at the mesoscopic scale are still so computational demanding that they are out of the reach of atomistic simulations .
this predicament can be ameliorated by adopting a coarse - grained description of the molecular degrees of freedom that effectively includes hydrodynamic modes .
one widely - used method is the lattice boltzmann scheme that simulates the evolution of the boltzmann equation for a simple fluid on a regular lattice by a series of collision and propagation steps for probability density functions defined on the sites of a lattice @xcite .
this scheme has been generalized to lcs with successful results @xcite . however , lattice boltzmann schemes have the limitation that they do not include thermal fluctuations . a different approach that emphasize
the particle aspect is the dissipative particle dynamics ( dpd ) .
this is an off - lattice , particle - based method , where particles represents fluid elements and are subject to pair - wise additive forces , which conserve momentum locally and thus generate the navier - stokes hydrodynamics for simple fluids .
the dpd scheme has also been extended to lcs @xcite . whereas this method has been shown to reproduce equilibrium phase diagrams of lcs @xcite , we are not aware of any attempt at reproducing nematodynamic behavior , though the original version of dpd does reproduce the correct hydrodynamics of simple fluids @xcite .
more recently , malevanets and kapral @xcite introduced the stochastic rotation dynamics ( srd ) model .
this is also an off - lattice , particle - based model where each particle represents a small parcel of fluid .
the fluid evolves through a series of collisions and streaming steps that exactly conserve mass , linear momentum and energy , and additionally respect galilean symmetry .
thus , the correct hydrodynamic modes are generated . because srd is a particle - based method , fluctuations are naturally present .
the srd method has been applied to a variety of systems , from colloids @xcite to polymers @xcite , from the modeling of the solvent carrying hydrodynamic interactions in vescicle self - assembly @xcite to the flow - induced shape transition in red blood cells @xcite . in general the srd model
is useful whenever both thermal fluctuations and hydrodynamic modes are physically important . in the present work
we extend the srd model to anisotropic fluids , namely lcs .
this is primarily done by giving orientational degrees of freedom to the srd particles which obey the ( simplified ) equations of nematodynamics in the formulation of ericksen - leslie . for the sake of simplicity
we restrict ourselves to methods and results in two dimensions ( 2d ) .
we show below that the srd is amenable of studying both the equilibrium and nonequilibrium behavior of lcs .
this new extension of the srd model opens the door to the investigation of a wealth of lc phenomena occurring at the mesoscale .
this work is organized as follows . in sec .
[ sec : srd_lc ] we describe the theoretical background of the equations used in the present model , and the details of the numerical implementation .
we validate this model for nematic lcs by considering three study cases in sec .
[ sec : applications ] : i ) the isotropic - nematic phase transition ; ii ) the production and annihilation of topological charges , and iii ) lc rheology under shear flow . finally , we summarize and discuss our results in sec . [
sec : discussions ] .
we start by considering the standard srd model that has so far been used to describe fluids with isotropic interactions .
the system is composed of @xmath3 particles of mass @xmath4 with positions @xmath5 and velocities @xmath6 , where @xmath7 $ ] . the evolution in time
@xmath8 proceeds through a series of two steps : + ( i ) the free - streaming step @xmath9 where all positions are updated ; + ( ii ) the collision step @xmath10\ ] ] where all velocities are updated by rotating the fluctuating part of the velocity with respect to the center of mass velocity @xmath11 in eq .
( [ eq : srd - vel]-[eq : srd - cm ] ) the calculations are performed in a cell - wise fashion , that is , the system is divided with a regular grid , and @xmath12 is computed from the @xmath13 particles within the cell @xmath14 to which particle @xmath15 belongs .
the rotation ( or collision ) matrix @xmath16 is orthogonal , @xmath17 , where the superscript @xmath18 denotes transposition .
it rotates , independently in each cell , the fluctuating part of the velocity by an angle @xmath19 about an arbitrary axis . because of its action on the fluctuating part of the velocity the collision rule conserves momentum ; because of its orthogonality the energy is also conserved . from these facts
follows that the fluid obeys the navier stokes equations . malevanets and
kapral showed @xcite that if @xmath16 satisfies detailed balance then the fluid approaches a maxwell boltzmann distribution and obeys the @xmath20-theorem . in practice
there are many ways to choose @xmath16 so that all the requirements are met . in 2d @xmath16
can only rotate the velocities by an angle @xmath21 with equal probabilities ; in 3d it is common , _
e.g. _ , to choose a random axis and rotate around it of a fixed angle . to ensure galilean invariance the grid
must be shifted by a random amount at each time - step @xcite so that artificial correlations among particles do not build up due to repeated collisions with the same neighbors .
we now come to the extension to nematic lcs .
the particle s degrees of freedom must be augmented by a unit vector describing its orientation @xmath22 , @xmath23 .
two particles separated by a distance @xmath24 interact through the lebwohl - lasher potential @xcite @xmath25 where @xmath26 indicates that @xmath15 and @xmath27 are neighbors .
lin _ et al . _
@xcite proposed a simplified version of the ericksen - leslie equations .
these are nonparabolic , dissipative equations that describe the flow of nematics in the one - constant approximation .
although they represent a drastic simplification of the original ericksen - leslie model , they retain the essential mathematical features , and they obey an energy law similar to the one used in the ericksen - leslie model @xcite .
we propose the following hybrid model of srd for anisotropic fluids @xmath28 where @xmath29 is the density , @xmath30 the pressure , @xmath31 the kinematic viscosity , and @xmath32 the ericksen - leslie stress tensor , that is the tensor whose @xmath33 component is @xmath34 . in eq . , @xmath35 is the elastic relaxation constant , the term @xmath36 is the molecular field inducing nematic ordering , @xmath37 its strength , and @xmath38 is a gaussian white noise for the director s angular velocity , with @xmath39 and @xmath40 , where @xmath41 is boltzmann s constant , @xmath42 is the temperature , @xmath43 is a kronecker delta , and @xmath44 is dirac s delta distribution . the ericksen - leslie stress tensor @xmath45 represents the feedback of the director field @xmath46 to the bulk flow of molecules . in this formulation
, @xmath45 is directly responsible for the non - newtonian behavior of the lc flow . because of their molecular anisotropy
the viscosity of a lc fluid does not depend solely on the shear stress @xmath47 , but the director field @xmath46 also participates in the viscosity generation .
this means that the velocity field can couple to the director and reorient it , and , also , that a reorientation of the director may generate a flow , usually called _
backflow_. we now describe the numerical implementation of our hybrid model for lcs .
we will make a distinction between quantities computed on a particle level , such as @xmath48 , and on a cell level , such as @xmath12 .
we note that the standard srd steps described in eqs .
( [ eq : srd - pos]-[eq : srd - vel ] ) recover the navier stokes equation , that is eq . without the last term .
also , eq . ( [ el - director ] ) applies in the eulerian picture ( lab - frame ) to the fluid parcel moving along the streamline . since in simulations
the director of the @xmath15th particle , @xmath22 , moves along the flow , the convective term @xmath49 is absorbed in this lagrangian picture ( comoving frame ) .
the srd algorithm for lcs consists of the following steps : + ( i ) free streaming @xmath50 this is identical to the standard srd step ; + ( ii ) cell - wise calculations , that is , the particles are grouped in different cells and @xmath12 , @xmath51 , gradients such as @xmath52 and @xmath53 , and the ericksen - leslie elasticity tensor @xmath45 are also calculated ; + ( iii ) lc alignment @xmath54\delta t\end{gathered}\ ] ] where eq .
is implemented .
+ ( iv ) collisions and backflow .
the cell - wise , center of mass velocity is calculated as in eq . , and the contribution from the ericksen - leslie tensor
is then added @xmath55\end{gathered}\ ] ] we consider here only a 2d system , thus the matrix @xmath16 rotates the particles thermal velocities around one axis , conventionally denoted as the @xmath56-axis , either clockwise or counter - clockwise by a fixed angle @xmath19 stochastically .
the parameter @xmath57 is the thermostat scaling factor whose role will be explained below . in general , srd
does not conserve angular momentum . to impose the conservation of the fluid s angular momentum in 2d the angle @xmath19
must be chosen @xcite such that @xmath58 where @xmath59|_z\ , , \\
b&=\sum_{i=1}^{\mathcal{n}_{c_i}}\vec r_i \cdot ( \vec v_i-\vec u_{c_i})\,.\end{aligned}\ ] ] a thermostat for translational and rotational velocity is implemented to control the temperature of the system .
equipartition of the energy is applied to keep the same amount of energy in each degree of freedom .
the temperature is defined as @xmath60 from the translational kinetic energy , or as @xmath61 from the rotational kinetic energy in 2d , where @xmath62 is the moment of inertia of the particles and @xmath63 is the angular velocity .
we employ a simple velocity rescaling thermostat that scales linear and angular velocities in a cell - wise fashion by a factor @xmath64 ( see eq . ) , where @xmath65 is the instantaneous kinetic temperature in cell @xmath14 .
we use two types of boundary conditions ( bc ) , periodic bc in both the @xmath66 and @xmath67 directions for simulations of bulk systems , and a no - slip wall perpendicular to the @xmath66 axis ( while periodic bc are implemented in the other direction ) for simulations of a channel geometry .
the no - slip bc are implemented with the usual bounce - back rule , that is , the velocity of the lc particle is inverted upon collision with the solid wall @xmath68 additionally , `` ghost '' particles @xcite are used at the walls to fill partially occupied cells up to the average particle density . at solid interfaces
the anchoring , that is , the preferential angle between lc particles and walls needs to be specified .
homeotropic anchoring is implemented by placing ghost particles in the walls with their orientations aligned perpendicularly to the walls .
the system is initialized by assigning the positions @xmath69 and velocities @xmath48 .
the particles are uniformly distributed in space .
the linear and angular velocities are assigned so that they are distributed according to a maxwell distribution at the target temperature , and that the equipartition theorem is obeyed .
we perform all our simulations with the following parameters : the srd rotation angle is @xmath70 , particle mass @xmath71 , moment of inertia @xmath72 , and the elastic and molecular relaxation constants are @xmath73 and @xmath74 .
it is a common choice to set the srd grid size @xmath75 and the timestep @xmath76 . at each time step the grid
is shifted by a random displacement vector with components uniformly distributed in the interval @xmath77 $ ] . in the following
we measure the temperature @xmath42 in units of @xmath78 , where @xmath79 is the maximum propagation speed of a particle ( related to the courant condition ) .
the mean free path @xmath80 is typically smaller than one but the grid - shift method avoids the build - up of spurious correlations .
we perform simulations of a system in equilibrium at fixed @xmath42 with periodic bc in both @xmath66 and @xmath67 directions . a useful way to characterize the degree of ordering of a nematic lc is to define an order parameter that is zero in the isotropic phase and one in the nematic phase . it is common to take the largest eigenvalue @xmath81 of the nematic order tensor @xmath82 which is a traceless , second - order tensor , and where @xmath83 is the dyadic product and @xmath84 is the unit tensor . . ]
we consider a system of @xmath85 particles , and of size @xmath86 , subdivided with a @xmath87 grid , thus the average number of particles per cell @xmath88 .
figure [ fig1 ] shows the dependence of the nematic order parameter @xmath81 on @xmath42 , at fixed density . at low @xmath42
the nematic order parameter is very close to one , indicating nearly perfect nematic alignment .
as @xmath42 increases @xmath81 decreases continuously until it reaches the constant value @xmath89 , which characterizes the isotropic phase in 2d @xcite .
we find that the isotropic - nematic transition occurs at @xmath90 .
a lc system in two dimensions or higher undergoes a temperature - driven isotropic - nematic phase transition .
this phase transition is continuous in 2d , as expected for a system with the symmetry of the xy model ( though there is some controversy @xcite ) .
contrary to the hydrodynamic ericksen - leslie model , our srd hybrid model can reproduce the isotropic - nematic phase transition .
the reason for that is the particle nature of our model .
different parts of the system may have different local orientations and thus , at high enough @xmath42 , the system is globally disordered . during a quench from the isotropic to the nematic phase a lc substance develops many topological defects @xcite which influence the kinetics of the phase transition .
solid boundaries imposing geometric or energetic ( anchoring ) constraints , or external fields in general may induce different , local orientations in a nematic lc .
because of the conflict between these local orientations topological defects are produced .
these are regions of the fluid where the local nematic order parameter is vanishing and the director field is undefined @xcite .
topological defect , while the red solid circle indicates the @xmath91 defect . ] ) .,title="fig : " ] ) .,title="fig : " ] although there were attempts of using ericksen - leslie model to simulate topological charges @xcite , those structures can not be properly treated since the nematic order parameter is essentially constant in that approach .
however in our model , the nematic order parameter in the macroscopic scale is calculated from the mesoscopic srd particle configurations , hence the formation of topological defects is an essential feature in this particle - based approach .
we investigate the generation of topological defects and their dynamics by quenching the system from a high value of @xmath42 , well in the isotropic phase , into the nematic phase at @xmath92 , with an equilibrium value of @xmath93 .
figure [ fig2 ] shows the temporal evolution of the topological defects formed during the quench . a large number of @xmath94 charges
are generated .
the total topological charge of the system is zero , and , by conservation of charge , an equal number of positive and negative charges are present .
immediately after the initial quench the srd particles start to align with their neighbors , due to the torque induced by the molecular field @xmath95 . after some time
the particles start to form larger nematic domains . at this stage ( fig .
[ fig2](a ) ) , there are many topological defects , carrying topological charges of @xmath96 and @xmath97 . at a later stage , those charges with opposite signs are attracted to each others and eventually annihilated with each others ( fig .
[ fig2](b - d ) ) , leaving the area charge - free as the initial configuration .
the speeds of topological defects depend on the charges they carry , that is , defects with charge @xmath96 move faster than @xmath97 charges ; this has been verified both by numerical @xcite and experimental work @xcite .
we find a similar behavior in our simulations .
figure [ fig3](a ) shows the time dependence of the positions of two topological charges .
initially , the charges approach each others with rather slow velocities , but when the two charges are very close they accelerate until they annihilate .
it is clear the velocity of the positive charge is higher than the negative one .
the reason for this asymmetry is that the velocity of the defect core depends on the sign of the spatial derivative of the director field @xmath52 .
similar result has been reported in @xcite where lattice boltzmann simulations of a tensorial formulation of nematodynamics were performed . to further verify the validity of our model we test the time dependence of the separation @xmath98 between opposite charges .
denniston predicted @xcite a simple scaling law @xmath99 where @xmath100 is the annihilation time and @xmath101 is a constant .
figure [ fig3](b ) shows that the separation between opposite charges does indeed follow the behavior predicted by eq . .
although the scaling law was derived in conditions of no backflow @xcite , it is still valid when the coupling of flow and director field is present , as observed in the tensorial treatments of nematodynamics @xcite
. then we conclude that our particle - based approach to nematodynamics can correctly describe regions of the fluid with strong gradients of the nematic order parameter , such as topological defects .
shear flow experiments are the canonical way to study the rheological properties of fluids .
the coupling of the elastic deformations of lcs with the transport and deformation of fluid elements gives rise to a more complex situation than in the flow of an isotropic fluid . in the original formulation of ericksen and leslie
@xcite the viscous response is characterized by six coefficients @xmath102 , @xmath103 , called leslie coefficients .
later , parodi showed @xcite that from onsager s reciprocity theorem follows @xmath104 , thus the number of independent viscosity coefficients is five .
our goal in this section is to verify if our hybrid model is capable of reproducing the known lc rheology . to generate a shear flow in a 2d simulation domain ,
we consider two no - slip walls at @xmath105 and @xmath106 and periodic bc in the @xmath67 direction . because of the presence of solid walls the boundary conditions for the director field need to be specified .
we employ homeotropic anchoring , that is , the lc particles prefer to orient perpendicularly to the walls .
this is achieved by assigning a perpendicular orientation of the ghost particles at the walls .
the two walls move with equal and opposite speeds in the @xmath67 direction . after some time
a stationary shear flow is generated .
a dimensionless measure of shear is given by the weissenberg number @xmath107 where @xmath108 is the shear rate and @xmath109 is a relaxation time .
we calculate @xmath109 from the orientational correlation function @xmath110 , where the angle brackets indicate ensemble average .
we present results for a fixed @xmath111 at @xmath112 ( we note that the presence of walls inducing homeotropic alignment shifts the phase diagram with respect to fig .
[ fig1 ] , so that @xmath112 corresponds to the nematic state ) .
the system of @xmath113 particles and size @xmath114 , @xmath115 is divided with a grid of @xmath116 , thus with @xmath117 .
the steady - state director fields and bulk flow profiles of sheared lc are shown in fig .
[ fig4 ] for different ericksen - leslie stress constant @xmath118 . for the case of @xmath119
the flow profile is a straight line as for a newtonian fluid . for higher values of @xmath118
we observe a shear banding effect . by
increasing @xmath118 the flow profile starts to have a kink near the walls .
shear banding is a nonequilbrium transition to a state where regions with different shear rates coexist , thus visible through different slopes of the velocity profile @xmath120 in fig .
nematic lcs can produce shear bands @xcite .
the existence of this phenomenon is a consequence of the competition between elastic energy and viscous dissipation , which in turn is caused by the feedback from lc orientation to the flow induced by @xmath45 . the local director field @xmath2 ( represented as black dashes in fig .
[ fig4](a)-(c))shows that while the director is aligned homeotropically at the walls , thus satisfying the bc , it is tilted in the central region of the channel .
this tilt angle is due to a well - known flow alignment mechanism in lcs .
the local nematic order map ( color code in fig . [ fig4](a)-(c ) ) shows a subtle change from a high degree of nematic alignment at the walls , induced by the anchoring conditions , to a slightly lower value in the central region of the channel .
we have introduced a new mesoscopic lc model . our model is based on the stochastic rotational dynamics scheme , which is a particle - based algorithm that ignores the computationally heavy molecular interactions but correctly generates the hydrodynamic modes described by the navier - stokes equations .
the model introduced here fills an important gap for mesoscopic simulation techniques of lcs .
thermal fluctuations are explicitly present , which are of fundamental importance in both thermodynamic and dynamic processes , but are neglected in other popular schemes , such as the lattice boltzmann approach .
furthermore , the mesoscopic scale and the hydrodynamic behavior is directly addressed .
we have shown that this model can be used to study various aspects of the physics of liquid crystals .
we have considered three study cases .
first , we have found that our model system undergoes an equilibrium phase transition from nematic to isotropic as the temperature increases .
the transition is found to be continuous , as it is expected to be in 2d .
second , we have studied the nonequilibrium dynamics of topological defects emerging in a quenched lc .
topological defects with @xmath121 charge are formed but quickly annihilate with each other .
the temporal dependence of the distance between two opposite charges is found to match remarkably well the theoretically predicted power law .
third , we have considered a shear flow situation and found that the lc system develops shear bands as the coupling parameter between flow and director reorientation is increased .
we conclude that the model captures the non - newtonian character of lc rheology .
this hybrid algorithm can be easily applied to complex geometries of the confining walls . for the sake of simplicity we restricted the present work to 2d .
however , any realistic implementation requires a 3d setup .
a generalization of the model to 3d is under way and will be discussed elsewhere .
59ifxundefined [ 1 ] ifx#1 ifnum [ 1 ] # 1firstoftwo secondoftwo ifx [ 1 ] # 1firstoftwo secondoftwo `` `` # 1''''@noop [ 0]secondoftwosanitize@url [ 0 ]
+ 12$12 & 12#1212_12%12@startlink[1]@endlink[0]@bib@innerbibempty @noop _
_ ( , ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) link:\doibase 10.1103/revmodphys.46.617 [ * * , ( ) ] @noop * * , ( ) link:\doibase 10.1039/c2sm27337c [ * * , ( ) ] link:\doibase 10.1039/c3sm50677k [ * * , ( ) ] link:\doibase 10.1080/02678292.2013.807939 [ * * , ( ) ] @noop * * , ( ) @noop * * , ( ) link:\doibase 10.1039/tf9332900883 [ * * , ( ) ] link:\doibase 10.1039/tf9332900945 [ * * , ( ) ] @noop * * , ( ) @noop ( ) @noop * * , ( ) \doibase http://dx.doi.org/10.1122/1.548864 [ * * , ( ) ] @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop _ _ ( , ) link:\doibase 10.1103/physreve.58.7475 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.61.2332 [ * * , ( ) ] @noop * * , ( ) http://stacks.iop.org/0953-8984/12/i=43/a=101 [ * * , ( ) ] link:\doibase 10.1103/physreve.67.061703 [ * * , ( ) ] link:\doibase 10.1098/rsta.2004.1414 [ * * , ( ) ] \doibase http://dx.doi.org/10.1063/1.1879852 [ * * , ( ) ] http://stacks.iop.org/0295-5075/19/i=3/a=001 [ * * , ( ) ] @noop * * , ( ) link:\doibase 10.1103/physrevlett.93.220601 [ * * , ( ) ] @noop * * , ( ) http://stacks.iop.org/0953-8984/16/i=38/a=012 [ * * , ( ) ] \doibase http://dx.doi.org/10.1063/1.2358983 [ * * , ( ) ] link:\doibase 10.1073/pnas.0504243102 [ * * , ( ) ] link:\doibase 10.1103/physreve.63.020201 [ * * , ( ) ] link:\doibase 10.1103/physreva.6.426 [ * * , ( ) ] link:\doibase 10.1002/cpa.3160420605 [ * * , ( ) ] link:\doibase 10.1002/cpa.3160480503 [ * * , ( ) ] _ _ , @noop ph.d . thesis , ( ) @noop * * , ( ) http://stacks.iop.org/0295-5075/56/i=3/a=319 [ * * , ( ) ] @noop @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop _ _ ( , ) @noop _ _ ( , ) @noop * * , ( ) link:\doibase 10.1007/s100510050168 [ * * , ( ) ] @noop * * , ( ) link:\doibase 10.1103/physreve.66.021712 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.95.097802 [ * * , ( ) ] @noop * * , ( ) link:\doibase 10.1021/ma970737h [ * * , ( ) ] link:\doibase 10.1007/s00397 - 008 - 0260 - 9 [ * * , ( ) ] | we introduce a new mesoscopic model for nematic liquid crystals ( lcs ) .
we extend the particle - based stochastic rotation dynamics method , which reproduces the navier - stokes equation , to anisotropic fluids by including a simplified ericksen - leslie formulation of nematodynamics .
we verify the applicability of this hybrid model by studying the equilibrium isotropic - nematic phase transition and nonequilibrium problems , such as the dynamics of topological defects , and the rheology of sheared lcs .
our simulation results show that this hybrid model captures many essential aspects of lc physics at the mesoscopic scale , while preserving microscopic thermal fluctuations . |
it has been known that nanostructures based on carbon possess unique mechanical , electrical and optical properties.@xcite among them particular attention has been given to graphene , a two - dimensional sheet of carbon atoms arranged in a hexagonal structure with low - energy electronic excitations described by the dirac equation.@xcite at the present time suspended graphene membranes up to @xmath0 m diameter are produced.@xcite this makes possible to investigate the van der waals and casimir interaction between graphene and different material structures , such as another sheet of graphene , atoms , molecules , dielectric and metallic plates , spheres etc .
respective theoretical investigations were performed using the phenomenological density - functional methods,@xcite second - order perturbation theory@xcite and , for multilayered carbon nanostructures , using the lifshitz theory.@xcite to apply the lifshitz theory , one needs the reflection coefficients on graphene over a wide frequency region .
it is customary to express the reflection coefficients in terms of the frequency - dependent dielectric permittivity , a concept which is not well defined for one - atom - thick carbon nanostructures . because of this , two models for the reflection coefficients on graphene with no use of dielectric permittivity were proposed , the hydrodynamic one@xcite and the dirac one.@xcite in the framework of the hydrodynamic model , graphene is considered as an infinitesimally thin positively charged sheet , carrying a homogeneous fluid with some mass and negative charge densities .
the hydrodynamic model was applied for calculation of the casimir and casimir - polder interactions.@xcite in the framework of the dirac model , it is taken into account that for energies below a few ev the dispersion relation for quasiparticles in graphene is linear with respect to the momentum , whereas it is quadratic in the hydrodynamic model .
the reflection coefficients of graphene at zero and nonzero temperature were found in refs .
@xcite and @xcite , respectively .
some calculations of the casimir - polder graphene - atom interaction were performed at zero temperature using both the hydrodynamic and dirac models.@xcite it should be remembered , however , that both these models are only approximations .
as was already mentioned , the hydrodynamic model disregards the dirac character of charge carriers in graphene . as to the dirac model
, it extends the linear dispersion relation for quasiparticles to any energy , whereas this property applies only at low energies .
because of this , one can conclude that in calculations of the casimir and casimir - polder forces the dirac model of graphene should be applicable at large separations between the test bodies , whereas the hydrodynamic model might work at short separation distances .
there is also another approach to the application of the lifshitz theory to graphene . in this approach graphene
is characterized by a spatially nonlocal dielectric permittivity depending both on the frequency and the wave vector .
different authors express such a dielectric permittivity either through the polarizability of one graphene layer @xcite or in the random phase approximation.@xcite it is well known that thermal casimir effect is a subject of debate and there are different theoretical approaches to its description.@xcite for two graphene sheets in a nonretarded regime it was found@xcite that a relatively large thermal correction to the casimir force at room temperature arises at short separation distances of tens of nanometers .
this conclusion was qualitatively confirmed,@xcite using the dirac model of graphene , with an alternative explanation .
the reason for the origin of large thermal correction for graphene is that the contribution of all terms with nonzero matsubara frequencies at room temperature becomes small in comparison with the zero - frequency term even at short separations .
it was also found@xcite that in graphene - atom casimir - polder interaction the thermal effect depends crucially on the magnitude of a mass gap parameter in the dirac model .
specifically , for a nonzero gap there exists an interval of temperatures ( separations ) where the thermal correction remains small with increasing temperature ( separation ) .
the possibility of large thermal correction at short separations links the dirac model of graphene to the drude model used in the literature to describe the casimir effect between real metals.@xcite because of this , it is of much interest to investigate the thermal casimir effect in the interaction of graphene with real material bodies made of different materials . in this paper
, we calculate the free energy of the casimir interaction between a suspended graphene membrane described by the dirac model and dielectric ( silicon , sapphire ) or metallic ( au , ni ) plates . in so doing materials of the plate
are described by realistic dielectric permittivities taking into account the interband transitions of core electrons .
we demonstrate that , similar to graphene - atom interaction , the behavior of thermal correction crucially depends on the mass gap parameter @xmath1 of the dirac model .
note that although the dirac - type excitations in pristine graphene are gapless , the influence of electron - electron interaction , substrates , defects of structure , and other effects leads to a nonzero mass gap.@xcite specifically , we show that larger is the magnitude of mass gap parameter , wider is the separation ( temperature ) interval , where the thermal correction remains small with increasing separation ( temperature ) . for a metallic plate interacting with graphene , we perform all calculations using both the drude- and plasma - model approaches to the dielectric permittivity of metal . in the case when graphene described by the dirac model interacts with a metallic plate ( either nonmagnetic or magnetic ) the calculation results obtained using both approaches nearly coincide and do not depend on the magnetic properties .
in contrast to the case of two drude metals , the thermal correction for a graphene - metal interaction has the same sign as the interaction energy at zero temperature , i.e. , the magnitude of the casimir free energy increases with increasing temperature . note that all results obtained for a free energy in graphene - plate geometry can be reformulated as the casimir force between a material sphere and a graphene sheet using the proximity force approximation ( pfa).@xcite as was recently shown,@xcite the error arising from the use of pfa is less than the ratio of separation distance to sphere radius . in this paper
we also compare the predictions of the dirac model for the thermal casimir effect with respective predictions of the hydrodynamic model and discuss the application region of each .
specifically , it is shown that the hydrodynamic and dirac models of graphene lead to different results at short separations and to nearly coinciding results at large separations for the casimir free energy of graphene interacting at room temperature with a dielectric plate or with a nonmagnetic metallic plate described by the drude model . for a nonmagnetic metallic plate described by the plasma model or for a magnetic plate the predictions of the hydrodynamic and dirac models of graphene
are significantly different at all separations considered and can be discriminated experimentally . at large separations
the asymptotic expressions for the casimir free energy of graphene - plate interaction are derived and compared with the computational results .
the paper is organized as follows . in sec .
ii we present the reflection coefficients of the electromagnetic oscillations on graphene in the dirac and hydrodynamic models .
section iii is devoted to the thermal casimir interaction of graphene described by the dirac model with a dielectric plate made of silicon or sapphire .
similar results for graphene interacting with a metallic plate made of au and ni are presented in sec .
iv . in sec . v the theoretical predictions following from the dirac and hydrodynamic models are compared .
our conclusions and discussions are contained in sec .
the reflection coefficients on graphene in the dirac model were expressed in terms of the components of the polarization tensor in three - dimensional space - time in the following way@xcite @xmath19 here , the dimensionless components of the polarization tensor are defined as @xmath20 and trace stands for the sum of spatial components @xmath21 and @xmath22 . at nonzero temperature
the explicit expression for the polarization tensor in the dirac model with arbitrary mass gap parameter @xmath1 and chemical potential @xmath23 was found in ref .
@xcite . here
we consider the case of undoped graphene and put @xmath24 . then in terms of our dimensionless variables the result of ref .
@xcite for the 00-component of the polarization tensor takes the following equivalent form:@xcite @xmath25^{1/2 } } + \frac{8\alpha}{{\tilde{v}}_f^2}\int_{0}^{1}dx \label{eq4 } \\ & & ~\times \left\{\vphantom{\frac{{\tilde{\delta}}^2+\zeta_l^2x(1-x)}{\left[{\tilde{\delta}}^2 + x(1-x)f(\zeta_l , y)\right]^{1/2 } } } \frac{\tau}{2\pi}\ln\left[1 + 2\cos(2\pi lx)e^{-g(\tau,\zeta_l , y ) } + e^{-2g(\tau,\zeta_l , y)}\right ] -\frac{\zeta_l}{2}(1 - 2x ) \frac{\sin(2\pi lx)}{\cosh{g(\tau,\zeta_l , y)}+ \cos(2\pi lx ) } \right . \nonumber \\ & & ~ \left .
+ \frac{{\tilde{\delta}}^2+\zeta_l^2x(1-x)}{\left[{\tilde{\delta}}^2 + x(1-x)f(\zeta_l , y)\right]^{1/2}}\ , \frac{\cos(2\pi lx)+e^{-g(\tau,\zeta_l , y)}}{\cosh{g(\tau,\zeta_l , y)}+ \cos(2\pi lx ) } \right\}. \nonumber\end{aligned}\ ] ] in this expression , @xmath26 in the fine - structure constant ( @xmath27 is the electron charge ) , @xmath28 is the dimensionless mass gap parameter , the dimensionless fermi velocity is @xmath29 , and the dimensionless variable @xmath30 is defined as @xmath31 .
equation ( [ eq4 ] ) also contains two dimensionless functions defined by @xmath32^{1/2}. \nonumber\end{aligned}\ ] ] the result of ref .
@xcite for the trace of the polarization tensor in terms of our dimensionless variables is given by@xcite @xmath33 \int_{0}^{1}dx\frac{x(1-x)}{\left[{\tilde{\delta}}^2 + x(1-x)f(\zeta_l , y)\right]^{1/2 } } + \frac{8\alpha}{{\tilde{v}}_f^2}\int_{0}^{1}dx \label{eq6 } \\ & & ~\times \left\{\vphantom{\frac{{\tilde{\delta}}^2+\zeta_l^2x(1-x)}{\left[{\tilde{\delta}}^2 + x(1-x)f(\zeta_l , y)\right]^{1/2 } } } \frac{\tau}{2\pi}\ln\left[1 + 2\cos(2\pi lx)e^{-g(\tau,\zeta_l , y ) } + e^{-2g(\tau,\zeta_l , y)}\right ] \right .
\nonumber \\ & & ~ -\frac{\zeta_l(1 - 2{\tilde{v}}_f^2)}{2}(1 - 2x ) \frac{\sin(2\pi lx)}{\cosh{g(\tau,\zeta_l , y)}+ \cos(2\pi lx ) } \nonumber \\ & & ~ \left .
+ \frac{{\tilde{\delta}}^2+x(1-x)[(1-{\tilde{v}}_f^2)^2\zeta_l^2- { \tilde{v}}_f^4y^2]}{\left[{\tilde{\delta}}^2 + x(1-x)f(\zeta_l , y)\right]^{1/2}}\ , \frac{\cos(2\pi lx)+e^{-g(\tau,\zeta_l , y)}}{\cosh{g(\tau,\zeta_l , y)}+ \cos(2\pi lx ) } \right\}. \nonumber\end{aligned}\ ] ] the properties of the reflection coefficients ( [ eq2 ] ) with the polarization tensor ( [ eq4 ] ) and ( [ eq6 ] ) were studied previously.@xcite in the framework of the hydrodynamic model discussed in sec .
i the reflection coefficients on graphene take the more simple form@xcite @xmath34 here , the dimensionless characteristic wave number of the graphene is defined as @xmath35 , where the dimensional wave number is @xmath36 in eq .
( [ eq8 ] ) @xmath37 is the number of @xmath38-electrons per unit area , @xmath39 is the electron mass .
note that the parameter @xmath40 of the hydrodynamic model does not depend on temperature .
thus , the reflection coefficients ( [ eq7 ] ) and the free energy ( [ eq1 ] ) in the case of hydrodynamic model depend on the temperature only through the matsubara frequencies .
a different situation arises in the dirac model . here
, the reflection coefficients ( [ eq2 ] ) depend on @xmath3 not only through the matsubara frequencies , but also through the components of the polarization tensor .
this links the dirac model of graphene to the drude - model approach to the thermal casimir force between real metals ( because the relaxation parameter of the drude model and , thus , the reflection coefficients are the explicit functions of temperature ) . in this respect
the use of the hydrodynamic model of graphene in calculations of the casimir force is analogous to the plasma - model approach where the reflection coefficients depend on @xmath3 only through the matsubara frequencies . as to the reflection coefficient of electromagnetic oscillations on thick metallic plate ( semispace )
, they have the standard form@xcite @xmath41^{1/2}}{\varepsilon_l y+\left[y^2+\zeta_l^2(\varepsilon_l\mu_l- 1)\right]^{1/2 } } , \nonumber \\ & & r_{\rm te}^{(p)}(i\zeta_l , y)= \frac{\mu_l y-\left[y^2+\zeta_l^2(\varepsilon_l\mu_l- 1)\right]^{1/2}}{\mu_l y+\left[y^2+\zeta_l^2(\varepsilon_l\mu_l- 1)\right]^{1/2}}. \label{eq9}\end{aligned}\ ] ] here , both the dielectric permittivity @xmath42 and the magnetic permeability @xmath43 are calculated along the imaginary frequency axis .
here , we investigate dependences of the free energy as functions of temperature and separation in the casimir interaction of a suspended graphene sheet with the dielectric plate made of different dielectrics . as the most typical dielectric materials we consider silicon ( si ) and sapphire ( al@xmath44o@xmath45 ) which possess relatively large values of the static dielectric permittivity [ @xmath46 and 10.102 , respectively ] but quite different behaviors along the axis of imaginary frequencies ( for si it is caused by electronic polarization , whereas for al@xmath44o@xmath45 by both electronic and ionic polarizations ) .
we begin with the free energy of interaction of graphene described by the dirac model with si plate .
the dielectric permittivity of si along the imaginary frequency axis was obtained@xcite by means of the kramers - kronig relations from the tabulated optical data for the complex index of refraction.@xcite it is assumed to be temperature - independent .
computations were performed by eq .
( [ eq1 ] ) , where the reflection coefficients on graphene are given by eqs .
( [ eq2 ] ) , ( [ eq4 ] ) and ( [ eq6 ] ) , and on silicon by eq .
( [ eq9 ] ) with @xmath47 and @xmath48 specified above , over the temperature interval from 0 to 300k at two separation distances @xmath49 nm and @xmath50 m . for the mass gap parameter of the dirac model of graphene only the upper bound is known.@xcite for a suspended graphene we choose the realistic upper bound @xmath51ev . for a graphene deposited on substrate , @xmath1 can be several times larger.@xcite taking this upper bound into account , we perform all computations for @xmath52 , 0.05 , 0.01 , and 0ev .
the computational results for the free energy per unit area are presented in fig .
[ fg1 ] as functions of temperature , where the lines from top to bottom correspond to @xmath1 decreasing from 0.1ev to 0ev , ( a ) at the separation @xmath49 nm and ( b ) at @xmath50 m . as can be seen in fig .
[ fg1](a , b ) , for each mass gap @xmath53 there exists the temperature interval where the free energy remains nearly constant with increasing temperature .
the width of these intervals quickly decreases with decreasing @xmath1 .
thus , at @xmath49 nm and @xmath54ev the free energy is nearly constant up to @xmath55k . at the same separation , but with @xmath56ev , the same property holds only up to @xmath57k .
the widths of intervals , where the free energy remains nearly constant with increasing @xmath3 , are also narrowed with the increase of separation [ see fig . [ fg1](b ) ] .
note that computations performed for @xmath58ev lead to nearly the same numerical results as for @xmath59 [ the lowest lines in fig .
[ fg1](a , b ) ] .
figure [ fg1 ] suggests that within the temperature interval , where the free energy is nearly flat , the thermal correction to the casimir energy at zero temperature should be relatively small .
we confirm this conclusion by the direct computation of the thermal correction to the casimir energy defined as @xmath60 for the same values of parameters , as in fig .
[ fg1 ] .
the computational results for the thermal correction as a function of temperature are presented in fig .
[ fg2 ] at separations ( a ) @xmath49 nm and ( b ) @xmath50 m .
the lines labeled 1 , 2 , 3 , and 4 correspond to the values of mass gap parameter @xmath52 , 0.05 , 0.01 , and @xmath61ev , respectively . as expected , for each line with @xmath53 there is some interval where the thermal correction remains nearly zero .
these intervals are just the same where the casimir free energy in fig .
[ fg1 ] remains nearly flat .
it is interesting to note that the thermal corrections are monotonously decreasing functions of temperature and have the same negative sign as the free energy .
this contrasts with the drude model approach to the casimir force between real metals where the thermal correction over a wide temperature interval is positive making the free energy the nonmonotonous function of temperature.@xcite it is instructive also to compute the relative thermal correction to the casimir energy defined as @xmath62 the computational results are presented in fig .
[ fg3 ] in percent as functions of temperature at separations distances ( a ) @xmath49 nm and ( b ) @xmath50 m .
the lines 1 , 2 , 3 , and 4 are labeled in the same way as in fig .
as can be seen in fig .
[ fg3 ] , all the relative thermal corrections are monotonously increasing functions of temperature whose character depends crucially on the magnitude of a mass gap parameter @xmath1 .
for example , at @xmath49 nm [ fig .
[ fg3](a ) ] the relative thermal correction of line 1 ( @xmath54ev ) is nearly equal to zero below @xmath55k but achieves 120.44% at @xmath63k . to compare , at the same separation
the relative thermal correction of line 4 ( @xmath58ev ) , which is much larger than the corrections of lines 13 at low temperatures , achieves only 59.40% at @xmath63k .
it is seen that at room temperature the relative thermal correction in the casimir interaction of graphene with si is rather large even at relatively short separations . from fig .
[ fg3](b ) it follows that at @xmath50 m the relative thermal correction at room temperature exceeds 4000% , i.e. , the absolute thermal correction exceeds the casimir energy at zero temperature by a factor of 40 .
this makes graphene interesting for experimental investigation of thermal effects in the casimir force .
next we consider the interaction of graphene described by the dirac model with sapphire plate .
the dielectric permittivity of sapphire along the imaginary frequency axis is well described@xcite in the ninham - parsegian approximation and was already used@xcite in computations of the casimir force .
computations of the free energy as a function of temperature were performed in the same way as for silicon .
the computational results for @xmath49 nm are presented in fig .
[ fg4](a ) by the four lines from top to bottom for @xmath52 , 0.05 , 0.01 , and @xmath61ev , respectively .
similar to fig .
[ fg1](a ) , there are temperature intervals where the free energy remains nearly constant . for sapphire , however , the respective magnitudes of the free energy are smaller than for a silicon . skipping the computational results at @xmath50 m [ which are similar to those presented in fig .
[ fg1](b ) ] , we present in fig .
[ fg4](b ) the more detailed computational results for the free energy in the temperature interval from 0k to 200k , where the lines from top to bottom correspond to @xmath52 , 0.09 , 0.08 , 0.07 , 0.06 , and 0.05ev , respectively . keeping in mind that the exact value of @xmath1 is not known , computations of this kind can be useful for the determination of @xmath1 from the comparison between experiment and theory . for sapphire ,
the computational results for the absolute and relative thermal corrections are similar to those for silicon ( see figs .
[ fg2 ] and [ fg3 ] ) . because of this we do not present them here . here ,
we present the computational results for the interaction of graphene described by the dirac model with dielectric plates as a function of separation .
the same equations and dielectric functions , as in sec .
iiia , are used .
taking into account that the casimir free energy strongly depends on separation , we normalize the results obtained on the casimir energy per unit area between two parallel plates made of ideal metal @xmath64 in fig . [ fg5](a ) the quantity @xmath65 at @xmath66k is plotted as a function of separation over the region from 50 nm to @xmath67 m .
the lines from bottom to top correspond to the mass gap parameter equal to 0.1 , 0.05 , and @xmath68ev , respectively . in all cases
the magnitude of the free energy decreases monotonously with the increase of separation ( the increase of @xmath65 is fig .
[ fg5 ] is explained by the fact that @xmath69 decreases with separation faster than @xmath70 ) . in fig .
[ fg5](b ) similar results are shown at @xmath63k . here , to avoid an overlap of the lines , we consider a more narrow separation region from 50 nm to @xmath71 m .
the bottom and top lines correspond to @xmath52 and @xmath68ev , respectively .
as can be seen from fig .
[ fg5](b ) , at @xmath63k the dependence of the casimir free energy on the mass gap parameter of the dirac model is very weak .
moreover , at separations above 50 nm the magnitude of the free energy @xmath6 decreases with separation as @xmath72 , so that the quantity @xmath65 is nearly linear function of separation .
similar computations were performed for a graphene interacting with a sapphire plate . in fig .
[ fg6](a ) the computational results for @xmath65 as a function of separation at @xmath66k are presented , where the lines from bottom to top correspond to @xmath73 , 0.05 , and @xmath68ev , respectively .
these results are similar to those presented in fig .
[ fg5](a ) for a silicon plate . again , at @xmath66k the casimir free energy is strongly affected by the value of @xmath1 . to illustrate this in more detail , in fig .
[ fg6](b ) we plot @xmath65 over the separation region from 50 nm to @xmath74 m where the lines from bottom to top correspond to @xmath52 , 0.09 , 0.08 , 0.07 , 0.06 , and 0.05ev . at @xmath63k ,
similar to fig .
[ fg5](b ) , the dependence of the free energy on @xmath1 becomes very weak and one obtains @xmath65 increasing nearly linear with the increase of @xmath2 .
the asymptotic behavior of the casimir free energy at high temperature ( or , equivalently , at large separations ) can be obtained by considering the zero - frequency contribution to the lifshitz formula ( [ eq1 ] ) .
this corresponds to large values of the dimensionless parameter @xmath30 introduced after eq .
( [ eq4 ] ) . by putting @xmath18 , @xmath75 in eq .
( [ eq4 ] ) one arrives at @xmath76 , \nonumber\end{aligned}\ ] ] where @xmath77^{1/2}. \label{eq14}\ ] ] equation ( [ eq13 ] ) can be identically rearranged to the form @xmath78 . \label{eq15}\end{aligned}\ ] ] the application region of eq .
( [ eq23 ] ) depends on the specific values of parameters .
thus , for si at @xmath63k eq .
( [ eq23 ] ) leads to less than 1% errors in the values of the free energy , as compared with the results of numerical computations , at @xmath91 nm for graphene with @xmath54ev and at @xmath92 m for graphene with @xmath93ev . for si at @xmath66k and graphene with @xmath54ev eq .
( [ eq23 ] ) is not yet applicable at @xmath94 m and for graphene with @xmath93ev works well for @xmath95 m .
it is interesting to compare the asymptotic expression ( [ eq23 ] ) with other results obtained in the literature .
thus , using the nonlocal dielectric function in the random phase approximation , the free energy of graphene interacting with a dielectric substrate ( sio@xmath44 ) at large separations was found@xcite to decrease as @xmath96 .
this is not in accordance with the main term of our result ( [ eq23 ] ) which demonstrates the classical limit , as is expected at large @xmath2 .
note that another work@xcite models the dielectric properties of graphene by the drude - type function and arrives at the @xmath72 scaling for graphene - graphene interaction which satisfies the classical limit .
the case of graphene interacting with metallic plate is of special interest . as was mentioned in sec .
i , there are two different theoretical approaches to the description of casimir effect between real - metal plates .
the drude model approach takes into account the relaxation properties of conduction electrons . in the framework of this approach ,
the imaginary part of the dielectric permittivity of the drude model is used to extrapolate @xmath97 , obtained from the measured optical data , to zero frequency . by contrast , the plasma model approach disregards relaxation processes and extrapolates @xmath98 to zero frequency by means of the simple plasma model .
although the drude model approach may seem preferable , as it takes into account some really existing property of metals , the experimental situation more likely favors the plasma model approach . in a series of precise independent measurements performed by the two experimental groups@xcite
the drude model approach was excluded at a high confidence level ( several experiments@xcite also excluded the influence of free charge carriers , that are present in dielectric materials at room temperature , on the casimir force ) .
the two experiments that support the drude model approach@xcite are not independent measurements of the casimir force ; they are based on fitting procedures between measured data for the total force and theoretical predictions using hypothetical models for the electric contribution to it . here
, we show that in the interaction of graphene described by the dirac model with metallic plate the results obtained are not sensitive to the approach used ( either drude or plasma ) .
this is , however , not the case when graphene is described by the hydrodynamic model ( see sec .
numerical computations of the casimir free energy per unit area between graphene described by the dirac model and au plate were performed by using eqs .
( [ eq1 ] ) , ( [ eq2 ] ) , ( [ eq4])([eq6 ] ) and ( [ eq9 ] ) with @xmath47 . the dielectric permittivity of au along the imaginary frequency axis was described either by the generalized drude - like model with temperature - dependent relaxation parameter@xcite or by the generalized plasma - like model.@xcite these models use the six - oscillator approximation for the optical data extrapolated to zero frequency by means of simple drude and plasma models , respectively , with the plasma frequency @xmath99ev and the relaxation parameter at room temperature @xmath100ev . at lower @xmath3
the lower values of @xmath101 according to the standard theory of electron - phonon interaction have been used.@xcite the computational results using the drude- and plasma - model approaches are found to be indistinguishable .
thus , at @xmath66k the relative difference between the casimir free energies computed using both approaches achieves the maximum value of 0.02% at @xmath102 nm , does not depend on @xmath1 in the limits of our computational accuracy , and decreases with the increase of separation . at @xmath63k this difference achieves the maximum values of 0.06% at @xmath59 and 0.07% at @xmath54ev , and again decreases with increasing @xmath2 .
figure [ fg7 ] presents the casimir free energy per unit area as a function of temperature ( a ) for @xmath49 nm and ( b ) for @xmath50 m . it can be seen that fig .
[ fg7 ] demonstrates the same characteristic features , as figs .
[ fg1 ] and [ fg4 ] plotted for dielectric plates , but the magnitudes of the free energy for the case of metallic plate are larger .
the most important novel qualitative effect , which is found for both dielectrics and metals , is that for each nonzero @xmath1 the free energy is nearly unchanged with the increase of @xmath3 within some temperature interval . in fig .
[ fg8 ] we present the computational results for the absolute thermal correction to the casimir energy at zero temperature , defined in eq .
( [ eq10 ] ) , as a function of temperature .
the lines 1 , 2 , 3 , and 4 correspond to the values of the mass gap parameter @xmath52 , 0.05 , 0.01 , and @xmath61ev , respectively .
this figure is analogous to fig .
[ fg2 ] plotted for si .
it demonstrates that for a metallic plate the thermal correction in the graphene - plate geometry behaves qualitatively in the same way as for a dielectric plate , but with slightly larger magnitudes of the thermal correction . the computational results for the relative thermal correction , defined in eq .
( [ eq11 ] ) are presented in fig .
[ fg9 ] as a function of temperature .
the lines 1 , 2 , 3 , and 4 again correspond to the same respective @xmath1 , as in fig .
figure [ fg9 ] is analogous to fig .
[ fg3 ] plotted for a dielectric plate ( silicon ) . for a metallic plate at @xmath49 nm the relative thermal correction at room temperature appears only slightly larger than for a dielectric plate . at @xmath50 m at room temperature
the relative thermal correction for au is smaller than for si .
this is explained by different values of the casimir energy at zero temperature . keeping in mind that in most experiments on the casimir force the temperature is preserved constant and measurements are performed at different separation distances , here we present the computational results for a free energy of graphene - metal interaction as a function of separation . in fig .
[ fg10](a ) the free energy of graphene interacting with au plate normalized on the casimir energy between ideal metal planes ( [ eq12 ] ) is shown .
the three lines from bottom to top correspond to the values of mass gap parameter @xmath52 , 0.05 , and @xmath68ev , respectively .
the obtained values of the free energy are larger than for graphene interacting with si plate [ compare with fig .
[ fg5](a ) ] . at @xmath63k
the dependence of the computational results on the mass gap parameter becomes not so pronounced as in fig .
[ fg10](a ) .
this can be seen in fig .
[ fg10](b ) where the bottom and top lines correspond to @xmath54ev and @xmath68ev , respectively .
it is interesting to consider the casimir interaction of graphene with a ferromagnetic metal .
it was shown@xcite that at room temperature the ferromagnetic properties of real metals may influence the casimir force only through the contribution of the zero - frequency term of the lifshitz formula .
recently the gradient of the casimir force between a nonmagnetic au sphere and a magnetic metal ( ni ) plate has been measured.@xcite we have computed the casimir free energy per unit area between a graphene described by the dirac model and ni plate using the same formalism , as for an au plate .
the dielectric permittivity of ni along the imaginary frequency axis was found from the tabulated optical data@xcite extrapolated to zero frequency either by the drude or by the plasma model with the plasma frequency @xmath103ev and the relaxation parameter at room temperature @xmath104ev.@xcite the value of @xmath105 for the static magnetic permeability of ni has been used .
it was found that relative differences in the computational results for the free energy of graphene - ni interaction , when ni is described using the drude- and plasma - model approaches , are as small as computed above for the interaction of graphene with an au plate .
the influence of magnetic properties on the free energy was also shown to be negligibly small .
the relative difference between the free energies of graphene - ni and graphene - au interactions for graphene with @xmath54ev computed at @xmath63k is equal to 6% at @xmath49 nm and decreases to 1% at @xmath50 m .
it is less for smaller values of the mass gap parameter .
note that even these small differences are not due to magnetic properties of ni but due to different plasma frequency and optical properties of ni as compared to au .
now we derive the analytic expression for the casimir free energy of graphene described by the dirac model interacting with metallic plate at @xmath106 .
the contribution of the tm reflection coefficient for graphene interacting with dielectric plate was obtained in eq .
( [ eq22 ] ) . taking into account that for metallic materials
@xmath107 defined in eq .
( [ eq20 ] ) is equal to unity , one obtains from eq .
( [ eq22 ] ) @xmath108 . \label{eq24}\ ] ] calculating the integral with respect to @xmath13 and using eq .
( [ eq17 ] ) , we arrive at @xmath109 \label{eq25 } \\ & & ~~ \approx -\frac{k_bt}{16\pi
a^2}\zeta(3)\left[1- \frac{\pi\tilde{v}_f^2}{2\alpha\tau\ln\left(2\cosh \frac{\pi\tilde{\delta}}{\tau}\right)}\right ] , \nonumber\end{aligned}\ ] ] where @xmath110 is the riemann zeta function .
this result in the special case @xmath111 was obtained in ref .
note that for a dielectric plate considered in sec .
iiic the contribution of the tm mode was in fact equal to the total free energy because @xmath83 for dielectrics . for metals
this is in general not so ( see below ) .
now we are in a position to consider metallic plates made of nonmagnetic and magnetic metals described within both the drude and the plasma model approaches and in all cases find the high - temperature behavior of the total casimir free energy .
we begin with a nonmagnetic metal described by the drude model approach . in this case from eq .
( [ eq9 ] ) one obtains that @xmath83 and in accordance with eq .
( [ eq31 ] ) the te contribution to the free energy @xmath122 vanishes .
thus , for the plate made of a nonmagnetic drude metal the total free energy of graphene - metal interaction at high temperature is given by eq .
( [ eq25 ] ) .
next we consider a nonmagnetic metal described by the plasma - model approach . in this case from eq .
( [ eq9 ] ) we have @xmath123 where the parameter @xmath124 is defined as @xmath125 and @xmath126 is the effective penetration depth of electromagnetic oscillations into the metal . substituting eq .
( [ eq32 ] ) in eq .
( [ eq31 ] ) and integrating with respect to @xmath13 , one obtains @xmath127 the total asymptotic expression for the free energy at high temperature is given by the sum of ( [ eq25 ] ) and ( [ eq34 ] ) .
note that the contribution of the te mode ( [ eq34 ] ) is a negligibly small correction because @xmath128 .
we are coming now to the consideration of magnetic metals described by the drude model . from eq .
( [ eq9 ] ) it follows @xmath129 the substitution of this reflection coefficient in eq .
( [ eq31 ] ) results in @xmath130 this term is again negligibly small , as compared to @xmath131 , so that the total casimir free energy at high temperature is well described by eq .
( [ eq25 ] ) .
finally we consider the asymptotic expression for the free energy of graphene interacting with a magnetic metal described by the plasma model . in this case from eq .
( [ eq9 ] ) we get @xmath132 . \nonumber\end{aligned}\ ] ] this is similar to eq .
( [ eq32 ] ) with the replacement of @xmath124 for @xmath133 .
thus , instead of eq .
( [ eq34 ] ) , one obtains @xmath134 . \label{eq38}\end{aligned}\ ] ] the total asymptotic expression for the free energy at high temperature is given by the sum of eq .
( [ eq25 ] ) and negligibly small addition ( [ eq38 ] ) depending on the properties of magnetic metal .
the obtained analytic expressions were found in good agreement with the results of numerical computations within appropriate temperature ( separation ) intervals .
on this section we compare the computational results for the free energy of graphene - plate interaction obtained using two different models of graphene discussed in secs .
i and ii . we find separation regions where the predictions of both models are distinct and similar and compare respective asymptotic expressions for the free energy at high temperature ( large separations ) . keeping in mind the possibility to compare theoretical predictions for the casimir force with the experimental data , we calculate the free energy of graphene - plate interaction as a function of separation for both dielectric and metallic plates .
computations were performed using the lifshitz formula ( [ eq1 ] ) where the reflection coefficients ( [ eq17 ] ) for graphene in the framework of the hydrodynamic model were used .
the computational results for @xmath65 as a function of separation are presented in fig .
[ fg11 ] by the dashed lines ( a ) at @xmath66k and ( b ) at @xmath63k . in the same figure
the respective results for @xmath65 computed using the dirac model of graphene are reproduced from fig .
[ fg5](a , b ) by the solid lines . the solid lines in fig .
[ fg11](a ) from bottom to top correspond to @xmath52 , 0.05 , and @xmath68ev , respectively . note that in the scale used in fig .
[ fg11](b ) the two lines of fig .
[ fg5](b ) overlap .
they are shown as a single solid line in fig .
[ fg11](b ) . as is seen in fig .
[ fg11](a ) , at @xmath66k the hydrodynamic model of graphene predicts much larger magnitudes of the casimir free energy than the dirac model .
thus , at @xmath66k the predictions of the hydrodynamic model for @xmath70 at @xmath135 m is by factors of 36.0 and 6.7 larger than predictions of the dirac model with @xmath54ev and @xmath136ev , respectively . at @xmath137
m the respective factors are 97.9 and 4.2 . at @xmath63k
[ see fig .
[ fg11](b ) ] the predictions of the hydrodynamic model are larger than the predictions of the dirac model by the factors of 2.5 and 1.3 at @xmath135 m and @xmath137 m , respectively . as is seen in fig .
[ fg11](b ) , at @xmath63k , @xmath138 m the asymptotic regime of large @xmath30 is already achieved and the predictions of the hydrodynamic and dirac models almost coincide . at @xmath66k [ fig .
[ fg11](a ) ] the asymptotic regime of large @xmath30 is achieved at much larger separations than those shown in the figure .
now we compare the predictions of the hydrodynamic and dirac models of graphene interacting with a metallic plate .
all computations were performed using the same formalism as above .
we considered the plates made of a nonmagnetic metal au and a magnetic metal ni .
each of these metals was described either using the drude- or the plasma - model approach .
the computational results for the normalized free energy @xmath65 are presented in fig .
[ fg12 ] at @xmath63k as a function of separation . in this figure
, the solid line is reproduced from fig .
[ fg10](b ) [ note that in the scale of fig . [ fg12 ] the two solid lines in fig .
[ fg10](b ) overlap ] .
this line is obtained using the dirac model of graphene interacting with a metallic plate , be it au or ni ( see sec .
the dashed and dotted lines 1 , 2 , 3 , and 4 are obtained from computations using the hydrodynamic model of graphene .
the computational data shown by the dashed lines 1 and 3 were found for an au plate described by the plasma- and drude - model approaches , respectively .
the data shown by the dotted lines 2 and 4 were computed for a ni plate also described by the respective plasma- and drude - model approaches . as can be seen in fig .
[ fg12 ] , the free energy of graphene - metal interaction with graphene described by the hydrodynamic model strongly depends on the metal used ( magnetic or nonmagnetic ) and on the chosen approach to the description of dielectric properties . only for the nonmagnetic metal ( au ) described by the drude model ( line 3 )
the behavior of the free energy is qualitatively similar to the case of dielectric plate and nears the prediction of the dirac model at large separations [ compare with fig .
[ fg11](b ) ] . as an example , the casimir free energy of graphene - metal interaction computed using the hydrodynamic model of graphene at @xmath50 m , @xmath63k is larger than the same quantity computed using the dirac model by factors of 2.37 ( for an au plate described by the plasma model ) , 2.15 ( for a ni plate described by the plasma model ) , 1.93 ( for an au plate described by the drude model ) , and 1.55 ( for a ni plate described by the drude model ) .
this allows comparison between different theoretical predictions and experimental data .
it is interesting that for lines 13 the magnitude of @xmath6 predicted by the hydrodynamic model is always larger than for the predictions of the dirac model ( solid line ) . as to the line 4 ( ni described by the drude model ) , the prediction for @xmath70 from the hydrodynamic model becomes less than from the dirac model at @xmath139 m and remains so at larger separations .
thus , at @xmath94 m the ratio between the predictions of hydrodynamic and dirac models is equal to 0.42 . as was mentioned in sec .
iiic , at high temperature ( large separations ) the zero - frequency term of the lifshitz formula ( [ eq1 ] ) alone determines the casimir free energy .
the reflection coefficients on the graphene , described by the hydrodynamic model , at zero frequency follow from eq .
( [ eq7 ] ) @xmath140 taking into account that for a hydrodynamic model at @xmath63k the high - temperature regime starts at @xmath141 m , we find from eq .
( [ eq8 ] ) that at these separations @xmath142 . substituting eq .
( [ eq39 ] ) for the reflection coefficient @xmath143 into the zero - frequency term of the lifshitz formula ( [ eq1 ] ) , for the tm contribution to the free energy at high temperature one obtains @xmath144 . \label{eq40}\ ] ] for a dielectric plate , using eq .
( [ eq20 ] ) and integrating in eq .
( [ eq40 ] ) , we get @xmath145 as noted in sec .
iiic , for a dielectric plate the contribution of the te mode to the free energy vanishes .
thus , eq . ( [ eq41 ] ) provides the complete expression for the free energy of graphene - dielectric interaction at high temperature .
equation ( [ eq41 ] ) coincides with the first term in eq .
( [ eq23 ] ) obtained for graphene described by the dirac model [ remind that for the dirac model of graphene the asymptotic expression ( [ eq23 ] ) becomes applicable at much smaller separations at the same room temperature ; see fig .
[ fg11](b ) ] . for a metallic plate , replacing @xmath107 with unity , we arrive at @xmath146 this result coincides with the first term of the asymptotic expression ( [ eq25 ] ) obtained for graphene described by the dirac model .
equation ( [ eq42 ] ) provides the total asymptotic expression for the free energy only in the case when a nonmagnetic metal of the plate is described by the drude model ( see the dashed line 3 approaching the solid line in fig .
[ fg12 ] when the separation distance increases ) .
now we consider the contribution of the te mode to the free energy of graphene - metal casimir interaction . from eq .
( [ eq39 ] ) we have @xmath147 where @xmath148 takes the maximum value @xmath149 and decreases with further increase of separation .
then , for the contribution of the te mode to the free energy at high temperature one obtains @xmath150 . \label{eq44}\end{aligned}\ ] ] for a magnetic metal described by the plasma model we use the reflection coefficient ( [ eq37 ] ) and substitute it in eq .
( [ eq44 ] ) .
all calculations are similar to the case of nonmagnetic metal , but the quantity @xmath124 is replaced for @xmath133 .
thus , instead of eq .
( [ eq45 ] ) , we obtain @xmath162 } 1 - 2\left[\beta+2\delta\sqrt{\mu(0)}\right]\right .
\label{eq53 } \\ & & ~~\left .
+ \beta^2\left[1+\frac{\pi^2}{6\zeta(3)}\right]+ 2\delta\sqrt{\mu(0)}\left[\beta+\delta\sqrt{\mu(0)}\right ] \frac{\pi^2}{3\zeta(3)}\right\}. \nonumber\end{aligned}\ ] ] in this equation , however , it is impermissible to neglect by the quantity @xmath133 as compared to @xmath163 .
substituting the numerical values of constants to eq .
( [ eq53 ] ) , we find @xmath164 \right
. \nonumber \\ & & ~~\left .
+ 2.368\beta^2 + 5.474\delta\sqrt{\mu(0)}\left[\beta+\delta\sqrt{\mu(0)}\right ] \right\}. \label{eq54}\end{aligned}\ ] ] by combining eq .
( [ eq54 ] ) and eq .
( [ eq42 ] ) , we arrive at the total free energy of graphene interacting with a magnetic metal described by the plasma model @xmath165 \right\}. \label{eq55}\end{aligned}\ ] ] the main contribution to eq .
( [ eq55 ] ) is by a multiple two larger than the main contribution to eq .
( [ eq25 ] ) obtained for the dirac model of graphene ( compare the dotted line labeled 2 and the solid line in fig .
[ fg12 ] ) .
the obtained analytic asymptotic expressions for the free energy of graphene described using the hydrodynamic model in graphene - metallic plate geometry is in good agreement with the results of numerical computations .
thus , for an au plate described by the drude- and plasma - model approaches at @xmath94 m , @xmath63k , the results of analytic and numerical calculations differ by 0.35% and 1.4% , respectively .
for ni the same relative differences are equal to 9.3% and 3% .
note that relatively large deviation obtained for ni plate described by the drude - model approach is explained by the fact that in this case at @xmath94 m the linear asymptotic regime is not yet achieved ( see the dotted line labeled 4 in fig .
[ fg12 ] ) . to conclude , we emphasize that all results for the casimir free energy obtained in this and previous sections are simply convertable to the casimir force @xmath166 in the experimentally relevant configuration of a sphere above a plate used in most of experiments on measuring the casimir force .
this can be done by means of the pfa which states that @xmath167 where @xmath168 is the radius of the sphere . in our case , where @xmath169 is the casimir free energy between a graphene sheet and a material plate , the force @xmath166 defined in eq .
( [ eq56 ] ) can be considered as the casimir force acting between a graphene sheet and a material sphere of radius @xmath168 . as was mentioned in sec .
i , the error introduced by the use of the pfa was recently proved@xcite to be smaller than @xmath170 ( i.e. or order of 0.1% for the typical values of parameters ) . keeping in mind that many effects considered above far exceed 100% , the use of the pfa in the comparison between experiment and theory is fully justified .
in the foregoing , we have investigated the casimir free energy and the thermal correction to the casimir energy at zero temperature for a suspended graphene sheet interacting with a material plate , either dielectric or metallic . in so doing graphene
was described by the fully relativistic dirac model with temperature - dependent polarization tensor .
the dielectric properties of the plate were described by the frequency - dependent dielectric permittivity taking into account the interband transitions of core electrons . for a metallic plate both the drude- and plasma - model approaches suggested in the literature have been used .
the main novel result obtained for both dielectric and metallic plates is that for graphene with any nonzero mass gap parameter @xmath1 there exists temperature interval where the casimir free energy remains nearly constant .
this happens under the condition @xmath171 , which should be satisfied with a large safety margin .
if this condition is satisfied , the thermal correction to the casimir energy at zero temperature remains negligibly small .
we have also demonstrated that under the condition @xmath172 the thermal correction becomes relatively large .
this makes possible large thermal corrections for a graphene sheet interacting with material plate at rather low temperature ( short separations ) . with respect to the interaction with a metallic plate ,
it was shown that for graphene described by the dirac model the computational results for the free energy are nearly independent on whether the drude- or plasma - model approach to the dielectric permittivity of metal is used . to a large extent
the free energy of graphene interacting with metallic plate is also independent on whether metal is nonmagnetic or magnetic if graphene is described by the dirac model . in all cases considered ( dielectric or metallic plate , nonmagnetic or magnetic , described by the drude- or plasma - model approach ) the analytic asymptotic expressions for the casimir free energy at high temperature ( large separations )
have been obtained and compared with the results of numerical computations .
the casimir free energies obtained using the dirac model of graphene were compared with those calculated using the hydrodynamic model .
it was shown that at moderate temperatures ( separations ) the magnitudes of the free energy computed using the hydrodynamic model of graphene differ significantly from that computed using the dirac model .
this can be used for the experimental test of these models . at large separations ( high temeperature )
the theoretical predictions from both models of graphene nearly coincide for a dielectric plate and for a nonmagnetic metallic plate described by the drude model . for a nonmagnetic metallic plate described by the plasma model and for a magnetic plate described by any model the hydrodynamic and dirac descriptions of graphene lead to quite different results for the free energy at large separations ( high temperature ) .
this fact can be also used for the experimental test of different models . in this respect
the investigation of the interaction between graphene and metamaterials@xcite is also of large interest .
we have also found analytic asymptotic expressions for the free energy at high temperature ( large separations ) when graphene described by the hydrodynamic model interacts with a dielectric plate or with a plate made of a nonmagnetic or magnetic metal . in the last two cases
both the drude- and plasma - model approaches have been used for a description of the dielectric properties of metal .
the calculation results obtained from the asymptotic expressions were found in a very good agreement with the results of numerical computations .
this work was supported by the dfg grant bo 1112/211 .
g.l.k . and v.m.m . are grateful to the institute for theoretical physics , leipzig university , where this work was performed , for kind hospitality . 99 m. s. dresselhaus , physica status solidi ( b ) * 248 * , 1566 ( 2011 ) .
a. h. castro neto , f. guinea , n. m. r. peres , k. s. novoselov , and a. k. geim , rev .
phys . * 81 * , 109 ( 2009 ) .
b. alemn , w. regan , s. aloni , v. altoe , n. alem , c. girit , b. geng , l. maserati , m. crommie , f. wang , and a. zettl , acs nano * 4 * , 4762 ( 2010 ) .
a. bogicevic , s. ovesson , p. hyldgaard , b. i. lundqvist , h. brune , and d. r. jennison , phys .
* 85 * , 1910 ( 2000 ) .
e. hult , p. hyldgaard , j. rossmeisl , and b. i. lundqvist , phys .
b * 64 * , 195414 ( 2001 ) .
j. jung , p. garca - gonzlez , j. f. dobson , and r. w. godby , phys . rev .
b * 70 * , 205107 ( 2004 )
. j. f. dobson , a. white , and a. rubio , phys .
lett . * 96 * , 073201 ( 2006 ) .
a. n. rudenko , f. j. keil , m. i. katsnelson , and a. i. lichtenstein , phys .
b * 84 * , 085438 ( 2011 )
. i. v. bondarev and ph .
lambin , phys .
b * 70 * , 035407 ( 2004 ) .
e. v. blagov , g. l. klimchitskaya , and v. m. mostepanenko , phys .
b * 71 * , 235401 ( 2005 ) .
g. barton , j. phys .
a * 37 * , 1011 ( 2004 ) .
g. barton , j. phys .
a * 38 * , 2997 ( 2005 ) .
g. w. semenoff , phys .
* 53 * , 2449 ( 1984 ) .
d. p. divincenzo and e. j. mele , phys .
b * 29 * , 1685 ( 1984 ) .
m. bordag , j. phys .
a : math . gen . * 39 * , 6173 ( 2006 ) .
m. bordag , b. geyer , g. l. klimchitskaya , and v. m. mostepanenko , phys .
b * 74 * , 205431 ( 2006 )
. e. v. blagov , g. l. klimchitskaya , and v. m. mostepanenko , phys .
b * 75 * , 235413 ( 2007 ) .
m. bordag , i. v. fialkovsky , d. m. gitman , and d. v. vassilevich , phys .
b * 80 * , 245406 ( 2009 ) .
i. v. fialkovsky , v. n. marachevsky , and d. v. vassilevich , phys . rev .
b * 84 * , 035446 ( 2011 )
v. churkin , a. b. fedortsov , g. l. klimchitskaya , and v. a. yurova , phys .
b * 82 * , 165433 ( 2010 )
. t. e. judd , r. g. scott , a. m. martin , b. kaczmarek , and t. m. fromhold , new . j. phys .
* 13 * , 083020 ( 2011 ) .
b. e. sernelius , europhys .
95 * , 57003 ( 2011 ) .
j. sarabadani , a. naji , r. asgari , and r. podgornik , phys .
b * 84 * , 155407 ( 2011 ) . v. b. bezerra , r. s. decca , e. fischbach , b. geyer , g. l. klimchitskaya , d. e. krause , d. lpez , v. m. mostepanenko , and c. romero , phys . rev .
e * 73 * , 028101 ( 2006 ) .
g. l. klimchitskaya , u. mohideen , and v. m. mostepanenko , rev .
phys . * 81 * , 1827 ( 2009 ) .
j. s. hye , i. brevik , j. b. aarseth , and k. a. milton , j. phys .
gen . * 39 * , 6031 ( 2006 ) .
g. gmez - santos , phys .
b * 80 * , 245424 ( 2009 ) .
m. chaichian , g. l. klimchitskaya , v. m. mostepanenko , and a. tureanu , phys .
a * 86 * , 012515 ( 2012 ) .
s. a. jafari , j. phys .
: cond . mat .
* 24 * , 205802 ( 2012 ) .
k. pyatkovskiy , j. phys .
: cond . mat .
* 21 * , 025506 ( 2009 ) .
v. p. gusynin , s. g. sharapov , and j. p. carbotte , new j. phys . * 11 * , 095013 ( 2009 ) .
gusynin and s. g. sharapov , phys .
b * 73 * , 245411 ( 2006 ) .
m. bordag , g. l. klimchitskaya , u. mohideen , and v. m. mostepanenko , _ advances in the casimir effect _ ( oxford university press , oxford , 2009 ) .
g. bimonte , t. emig , r. l. jaffe , and m. kardar , europhys . lett .
* 97 * , 50001 ( 2012 ) .
l. p. teo , m. bordag , and v. nikolaev , phys .
d * 84 * , 125037 ( 2011 ) .
g. bimonte , t. emig , and m. kardar , appl .
. lett . * 100 * , 074110 ( 2012 ) .
a. o. caride , g. l. klimchitskaya , v. m. mostepanenko , and s. i. zanette , phys .
a * 71 * , 042901 ( 2005 ) .
b. geyer , g. l. klimchitskaya , and v. m. mostepanenko , phys .
a * 72 * , 022111 ( 2005 ) . ,
vol ii , ed .
e. d. palik ( academic , new york , 1991 ) .
l. bergstrm , adv .
colloid interface sci .
* 70 * , 125 ( 1997 ) .
d. drosdoff and l. m. woods , phys .
b * 82 * , 155459 ( 2010 )
. r. s. decca , e. fischbach , g. l. klimchitskaya , d. e. krause , d. lpez , and v. m. mostepanenko , phys .
d * 68 * , 116003 ( 2003 ) .
r. s. decca , d. lpez , e. fischbach , g. l. klimchitskaya , d. e. krause , and v. m. mostepanenko , ann .
( n.y . ) * 318 * , 37 ( 2005 ) .
r. s. decca , d. lpez , e. fischbach , g. l. klimchitskaya , d. e. krause , and v. m. mostepanenko , phys .
d * 75 * , 077101 ( 2007 ) .
r. s. decca , d. lpez , e. fischbach , g. l. klimchitskaya , d. e. krause , and v. m. mostepanenko , eur .
j. c * 51 * , 963 ( 2007 ) .
chang , a. a. banishev , r. castillo - garza , g. l. klimchitskaya , v. m. mostepanenko , and u. mohideen , phys .
b * 85 * , 165443 ( 2012 ) .
f. chen , g. l. klimchitskaya , v. m. mostepanenko , and u. mohideen , optics express * 15 * , 4823 ( 2007 ) .
f. chen , g. l. klimchitskaya , v. m. mostepanenko , and u. mohideen , phys .
b * 76 * , 035338 ( 2007 ) .
j. m. obrecht , r. j. wild , m. antezza , l. p. pitaevskii , s. stringari , and e. a. cornell , phys . rev
. lett . * 98 * , 063201 ( 2007 ) .
g. l. klimchitskaya and v. m. mostepanenko , j. phys . a : math
. theor . * 41 * , 312002(f ) ( 2008 ) .
chang , a. a. banishev , g. l. klimchitskaya , v. m. mostepanenko , and u. mohideen , phys .
lett . * 107 * , 090403 ( 2011 ) .
a. a. banishev , c .- c .
chang , r. castillo - garza , g. l. klimchitskaya , v. m. mostepanenko , and u. mohideen , phys .
b * 85 * , 045436 ( 2012 ) .
a. o. sushkov , w. j. kim , d. a. r. dalvit , and s. k. lamoreaux , nature phys .
* 7 * , 230 ( 2011 ) .
d. garcia - sanches , k. y. fong , h. bhaskaran , s. lamoreaux , and h. x. tang , phys .
lett . * 109 * , 027202 ( 2012 ) .
g. bimonte , phys .
a * 81 * , 062501 ( 2010 ) .
r. zandi , t. emig , and u. mohideen , phys .
b * 81 * , 195423 ( 2010 ) . v. b. bezerra , g. l. klimchitskaya , v. m. mostepanenko , and c. romero , phys . rev .
a * 69 * , 022119 ( 2004 ) .
b. geyer , g. l. klimchitskaya , and v. m. mostepanenko , phys .
b * 81 * , 104101 ( 2010 ) .
g. l. klimchitskaya , b. geyer , and v. m. mostepanenko , int . j. mod
a * 25 * , 2293 ( 2010 ) .
a. a. banishev , c .- c .
chang , g. l. klimchitskaya , v. m. mostepanenko , and u. mohideen , phys .
b * 85 * , 195422 ( 2012 ) , vol i , ed
. e. d. palik ( academic , new york , 1985 ) .
m. a. ordal , r. j. bell , r. w. alexander jr .
, l. l. long , and m. r. querry , appl
* 24 * , 4493 ( 1985 ) .
v. n. marachevsky , int . j. mod
* 14 * , 435 ( 2012 ) .
d. drosdoff and l. m. woods , phys .
a * 84 * , 062501 ( 2011 ) . | we investigate the thermal casimir interaction of a suspended graphene described by the dirac model with a plate made of dielectric or metallic materials .
the reflection coefficients on graphene expressed in terms of a temperature - dependent polarization tensor are used .
we demonstrate that for a graphene with nonzero mass gap parameter the casimir free energy remains nearly constant ( and the thermal correction negligibly small ) over some temperature interval . for the interaction of graphene with metallic plate ,
the free energy is nearly the same , irrespective of whether the metal is nonmagnetic or magnetic and whether it is described using the drude- or plasma - model approaches .
the free energy computed using the dirac model was compared with that computed using the hydrodynamic model of graphene and big differences accessible for experimental observation have been found . for dielectric and nonmagnetic metallic plates described by the drude model
these differences vanish with increasing temperature ( separation ) .
however , for nonmagnetic metals described by the plasma model and for magnetic metals , a severe dependence on the chosen theoretical description of graphene remains even at high temperature . in all cases the analytic asymptotic expressions for the free energy at high temperature
are obtained and found in a very good agreement with the results of numerical computations . |
the first generation of gravitational wave ( gw ) interferometric detectors has reached an unprecedented sensitivity to gw signals .
six interferometric detectors completed commissioning activities and acquired scientific data over the last years .
three of them constitute the laser interferometric gravitational - wave observatory ( ligo ) , a joint caltech - mit project supported by the national science foundation @xcite .
they are situated in the usa , one in livingston , louisiana ( l1 ) and two , which share the same facilities , in hanford , washington ( h1 , h2 ) . l1 and h1 are interferometers with arms of length 4 km , whereas h2 is a 2 km interferometer .
their first science run began in september 2002 .
since then , four other science runs took place ; between november 2005 and september 2007 they operated at their design sensitivity in a continuous data - taking mode in the 5th science run ( s5 ) .
the virgo detector is a joint project of the cnrs and infn , operated by the virgo collaboration at the european gravitational observatory - a cnrs - infn joint venture with the mandate to build the interferometer @xcite .
virgo is a 3 km interferometer near pisa ( italy ) .
its first science run ( vsr1 ) took place in may 2007september 2007 , with a sensitivity comparable to ligo instruments above around 400 @xmath0 and a better sensitivity below 30 @xmath0 .
( see fig .
[ fig : shh ] ) . typical spectral density of calibrated noise for the ligo interferometers and geo during the s5 run and for virgo interferometer during the vsr1 run . also displayed
are the design sensitivities for the 4 km ligo interferometers and for virgo.,width=377 ] two additional interferometric detectors are part of the network of gw observatories : the japanese tama project built a 300 m interferometer near tokyo , japan @xcite and the german - british geo project built a 600 m interferometer near hannover , germany @xcite ; early in its operation geo joined with the ligo scientific collaboration .
the current sensitivity of tama and geo is not comparable to ligo and virgo , and the latest scientific observational results are based on the analysis of s5/vsr1 data .
the ligo s5 run collected a full year of triple detector coincidence interferometer data , whereas virgo ran in science mode for about 110 days during vsr1 .
one of the most promising candidates for a gw detection is thought to be the signal originated during the coalescence of two neutron stars .
the distance at which a system of two neutron stars of 1.4 solar masses can be detected by a gravitational wave observatory , with a signal to noise ratio of 8 , has thus become a standard measure of the detectors sensitivity . during the s5/vsr1 run
, h1 reached a maximum sight distance of 35 mpc , corresponding to a spherical observable volume of radius of 15 mpc .
ligo and virgo signed an agreement to jointly analyze data from the four ligo - virgo interferometers collected after may 2007 .
the data analysis activity is run by four ligo - virgo search physics groups with different scientific targets : burst , compact binary coalescences , continuous waves and stochastic background , each aiming at the detection and characterisation of different sources of gws .
even if this division is somehow arbitrary , it has become the accepted standard in gw data analysis .
currently , a new joint run , s6/vsr2 , is on - going .
the goal is to improve the sensitivity of the detectors , enhanced ligo and virgo+ , by about a factor 2 over the course of the run , resulting in an observable volume of the universe of about an order of magnitude larger . in the next sections of this article a general introduction to the methods applied and observational results obtained for the different targets will be presented .
gravitational wave bursts are taken to be short signals , duration @xmath1 1s , with a large variety of possible waveforms , either wide or rather narrow - band in the frequency domain .
gw bursts historically correspond to poorly modelled signals , due to the complex physics involved in the description of the sources , such as core collapse supernovae or mergers of two compact stars ( black holes and/or neutron stars ) .
hence analysis methods seeking gw bursts must in general stay robust with respect to the expected waveforms .
moreover the benefit of running robust analysis methods is to stay wide open to the unexpected , something which is fine when in a discovery mode .
this constraint can be relaxed somewhat if some additional information about the source is available from other astrophysical observations .
also , spectacular recent progress in numerical relativity has allowed for accurate modelling of gw sources . for the most important burst sources we have now at least an idea of what the emitted waveform looks like . core collapse supernova
the collapse of a massive star and neutron star formation has long been envisaged as one of the first gw sources .
modern models now include 3d full relativistic hydrodynamical simulations together with detailed micro - physics @xcite .
they predict the emission of gw signal during the fast gravitational collapse and bounce with a duration @xmath2 ms and peak amplitude @xmath3 for sources located at 10 kpc . a longer ( hundreds of ms ) and in some cases stronger
signal is also expected after the bounce due to hydrodynamical instabilities around the proto - neutron star .
another source of gw bursts is the merger of two compact stars ( black hole and/or neutron star ) which generally results in a black hole , surrounded by a torus of matter if at least one of the progenitors is a neutron star .
progress in numerical relativity in recent years allows a clear picture of the merger of two black holes and a good prediction for the waveform @xcite .
good progress has been made too in modelling the merger of two neutron stars @xcite , which is an even more complex problem due to , for example , the role of the neutron star equation of state .
core collapse supernova and merger of neutron stars are thought to be the origin of , respectively , long and short gamma ray bursts ( grb ) . in both cases ,
a black hole will be formed and surrounded by a disk of matter .
the accretion of matter onto the black hole leads to the formation of relativistic jets and the emission of gamma rays .
what is interesting in this situation is that a gw signal and a grb signal are emitted within a short delay .
this may considerably enhance the detection capability , as explained below , and give deeper insight into the grb mechanism .
soft gamma repeaters ( sgr ) sporadically emit brief and intense bursts of soft gamma rays and could also be a good source of gw bursts .
sgrs are produced by highly magnetized neutron stars which undergo deformations ( starquakes ) that could excite the modes of the star and then emit gravitational waves at about the same time as the gamma emission .
the expected waveform is typically a damped sinusoid ( `` ringdown '' ) but the parameters depend on the equation of state of the nuclear matter inside the star and are not accurately known . like in the grb case
, the gamma ray emission permits a search for the gw bursts in a narrow coincidence time window .
based on the classification of expected sources given in the previous section , there are basically two ways of searching for gw bursts : a general one making very little or no assumption about the signal direction , waveform , time of arrival , etc . , and a second where we take benefit of an external trigger ( for example a grb ) for which source location in the sky and timing are known .
the main steps of the analysis are trigger generation , efficiency estimation and accidental background estimation , whose principles are the same for both all - sky all - time unmodelled searches and triggered searches .
even with increasingly accurate predictions for gw burst waveforms , it is still mandatory that burst algorithms remain as robust as possible against the possible variety of waveform . indeed , generic signals such as gaussian pulses or sine - gaussian signals are used in parallel with astrophysical waveforms ( collapse , mergers ) to test the efficiency and robustness of trigger generators .
the up - to - date trigger generator algorithms ( or analysis `` pipelines '' ) are time - frequency or equivalent methods looking at local excess of energy ( in time and frequency ) in the calibrated @xmath4 time series @xcite .
these methods can be fully coherent @xcite taking full benefit of the existing network of detectors ( ligo and virgo ) or coincident with some coherent follow - up when looking for possible event candidates .
a coherent method has the advantage that it can faithfully reconstruct the burst signal and the position of the source in the sky .
of course a coincident analysis can also reconstruct the sky location of some burst event candidate . in both cases
accurate sky position can be reconstructed if the three ligo and virgo interferometers are in operation and in science mode at the same time . in the case of the current ligo - virgo network ,
the typical position angular accuracy is of the order a few degrees , depending on the gw direction with respect to the detectors plane @xcite .
the efficiency estimation of the searches is performed by injecting signals in the calibrated time - shifted output data of the detectors .
as already mentioned , these signals can be astrophysical signals from core - collapse or merger simulations ( see for example the search for the merger and ring - down phases of binary black hole coalescences with virgo c7 commissioning run data @xcite ) or generic signals .
the so - called sine - gaussian signals ( sinusoids with gaussian envelope ) are particularly interesting since they can span the detector bandwidth and give sensitivity estimates for the whole range of accessible frequencies . a number of such signals
are then added to the calibrated strain of each detector after time sliding .
for the all sky searches , they are generally distributed along the entire data in order to average the detectors noise non - stationarities . for each injected
signal the amplitude described by the @xmath5 , which corresponds to the total gw signal energy : @xmath6 the @xmath5 amplitude has units of hz@xmath7 , meaning we can compare it directly to the detector sensitivities , especially when the injected waveform is well localized in frequency .
the final result for each signal is the efficiency curve , detection efficiency vs. @xmath5 .
usually the @xmath5 values at 50% or 90% efficiency are chosen for setting the upper limits .
another important aspect of the burst analysis is the background estimation and the rejection of loud outliers ( glitches ) using data quality flags and event by event vetoes .
the quality of the detectors data is studied at the individual detector level .
short duration noise transients can mimic a gw burst especially if their rate is large enough to induce a number of coincident events between the detectors .
data quality flags are used to indicate that one interferometer is not working properly during some period , for some known reason .
event - by - event vetoes are defined when an excess of coincidence is found between the gravitational strain channel triggers and auxiliary , environmental channel events .
the above triggers are generally not obtained with the search pipelines but by other algorithms which are required to be much faster since they run on hundreds of auxiliary channels .
the procedure allows us to veto only short time intervals ( sub - second ) , thereby saving the observation time of the detectors .
once the data is cleaned by use of data quality flags and vetoes it remains to estimate the background , i.e. , the number of accidental coincidences . as the noise in each detector is independent , this is done by time shifting the detector data with respect to the others and processing all these time shifted data sets with the search algorithms .
relative time shifts are chosen to be much larger than the light times of flight between the detectors and much larger than the expected signal durations .
there is a little difference for triggered searches . in this case , the data are split into an on - source region where the gw burst is expected and a background region where the noise is expected to be statistically similar to the one in the on - source region .
the background region generally consists of a few hours from each side of the on - source region . all the efficiency and background studies
are performed with the background data set .
all the cuts of the searches are then defined with these shifted data sets and applied in a blind way to the original unshifted data set .
events passing all the cuts ( if any ) are then detection candidates and can be examined more closely . if no candidate is found the result of the search can be turn into exclusion plots , for instance rate of event vs. @xmath5 in the case of all - sky analysis .
the frequentist upper limit for the rate of events with confidence level @xmath8 is for example @xcite @xmath9 where @xmath10 is the observation time and @xmath11 is the detection efficiency for the considered waveform with amplitude @xmath5 .
the search for gw bursts has not yet yielded a detection and most published results give upper limits on strain ( @xmath5 ) or rate .
when an astrophysical interpretation for a particular source is possible then these limits can be converted into meaningful astrophysical bounds .
the most recent all sky search for unmodelled gw bursts has been completed using data from the first calendar year of the s5 ligo run .
the analysis has been split in two parts , a search for low frequency signals in a 64 - 2000 hz band @xcite and a high frequency search in a 1 - 6khz band @xcite .
the observation time @xmath12 days sets an upper limit about 3.6 events / year for the rate of events at 90% confidence level ( @xmath13 in the formula above ) in the lower frequency band . in the higher frequency band
the network live - time is only @xmath14 days , resulting in an upper limit for the rates of events about 5.4 events / year at 90% confidence level .
the strain limits depend on the details of the injected signals for efficiency estimates .
the typical @xmath15 limit for sine - gaussian signals and most gaussian pulses is below @xmath16 hz@xmath7 in the low frequency band while it is up to a few @xmath17 hz@xmath7 in the high frequency band .
not surprisingly the search sensitivity follows the detector sensitivities which are better at low frequencies than at high frequencies .
another recent result concerns the search for a gw burst associated with grb 070201 @xcite , detected by gamma - ray satellites . those satellites
found that the error box for the position of the grb is centered at about 1 degree from the center of m31 ( andromeda ) and overlaps the spiral arms of the galaxy .
andromeda is the closest spiral galaxy ( at about 760 kpc ) and an event possibly occurring at such a short distance would be outstanding .
the grb was a short one and likely progenitors for short grbs are binary neutron star or neutron star - black hole coalescences or flares from sgr .
the analysis for coalescence waveforms is presented in the next section .
an unmodelled search for a gw burst in association with the grb , yielded an upper limit for the radiated gw energy of about @xmath18 m@xmath19c@xmath20 for gw bursts lasting less then 100 ms for isotropic emission occurring at the ligo peak sensitivity near 150hz .
this does not rule out the possibility of a sgr giant flare in andromeda galaxy .
other searches for gw bursts associated to grb have previously been pubblished by the ligo collaboration @xcite and virgo collaboration @xcite .
two searches for gw bursts possibly associated with sgrs have been recently published @xcite .
the first one targeted the sgr 1806 - 20 and the sgr 1900 + 14 , including a giant flare episode of sgr 1806 - 20 which occurred in december 2004 ( during the ligo `` astrowatch '' period prior to the s4 run ) and a storm episode of sgr 1900 + 14 which occurred in march 2006 ( during the ligo s5 run ) .
using a set of different waveforms the search was able to set upper limits at 90% confidence level on the isotropic gw emitted energies in the range @xmath21 to @xmath22 ergs for a source located at 10 kpc .
these upper limits depend on the detector sensitivities and antenna patterns at the time of the gamma emission , on the loudest event in the on - source region and the injected waveforms for efficiency estimation .
it is worth noting that some theoretical models predict maximal gw emission energy as high as @xmath23 ergs .
this is well in the range of the sgr analysis sensitivity .
the second paper is a re - analysis of the sgr 1900 + 14 storm of march 2006 with a different ( `` stacking '' ) method @xcite .
the gain in sensitivity is about one order of magnitude with respect to the first analysis @xcite .
a more exotic analysis is the search for gw bursts emitted by cosmic string cusps .
the first reported results used the 2005 ligo data ( s4 run ) @xcite .
the search used matched filtering as the predicted signals have simple and well parametrized waveforms .
upper - limits have been set for the rate of events and for cosmic string parameters such as string tension loop size or reconnection probability .
these limits are not competitive with the ones obtained by other cosmological observations like indirect bounds from big bang nucleosynthesis , but analysis of the s5 data ( much longer data taking and with sensitivity twice as good ) might surpass current limits in a some portion of the cosmic string parameter space .
coalescing binaries comprised of black holes and/or neutron stars are ideal sources for gravitational wave detectors . during the latter stages of its evolution ,
a binary emits gravitational radiation as the two component stars slowly spiral inwards before finally merging to form a single object which settles down to equilibrium .
indeed , the emission of gravitational waves from binary inspiral has been indirectly detected through observations of binary pulsars @xcite , although the current gravitational wave frequency is too low to be observed in terrestrial gravitational wave detectors .
as the inspiral progresses , however , the frequency and amplitude of the gravitational waves increase . during the final seconds or minutes of inspiral and merger , the gravitational radiation emitted by systems with a mass between one and several hundred solar masses will lie in the sensitive band of ground based gravitational wave detectors .
the precise form of the binary coalescence gravitational waveform depends sensitively upon the parameters of the binary , most notably the masses and spins of the binary components .
the eccentricity of the orbit will affect the emitted waveform but , in most cases , it is expected that the binary will have circularized before entering the sensitive band of ground based detectors @xcite .
historically , the binary coalescence has been split into three parts : a slow inspiral , a highly relativistic merger and ringdown to a final equilibrium state .
different techniques are used to calculate the waveform in each of these regimes . depending upon the mass of the system , different stages of the evolution
will emit gravitational waves at the sensitive frequency of the detector .
when the components of the binary are widely separated , the orbit decays slowly due to energy emitted in gravitational radiation , and the waveform sweeps slowly upwards in both frequency and amplitude . during this inspiral phase , theoretical waveforms calculated within the post - newtonian framework
@xcite are expected to provide an accurate representation of the gravitational waveform .
the post - newtonian waveforms derived to date are sufficient for binaries comprised of neutron stars or low mass black holes ( up to about @xmath24 ) , as the merger will occur above the most sensitive frequency band of the detector .
it is expected that the end product of a binary system with total mass above @xmath25 will be a single , perturbed black hole which rings down , by emission of gravitational radiation , to an equilibrium conguration .
since stationary black holes are fully characterized by their mass and angular momentum , the excitations of higher multipoles , and in partiular the quadrupole , will be radiated gravitationally @xcite .
the frequency and damping time for each of the ringdown modes depends upon the mass and angular momentum of the black hole and can be calculated analytically within the framework of black hole perturbation theory @xcite .
ringdown waveforms will lie in the sensitive band of the detector for black holes with mass greater than around @xmath26 .
the dynamical merger of two black holes can only be modelled using full general relativistic calculations .
recent breakthroughs in numerical relativity have , for the first time , enabled the calculation of the gravitational waveform emitted during merger @xcite .
currently , several groups are capable of numerically evolving two black holes through their final orbits , merger and ringdown @xcite .
the waveform for binaries whose components have comparable mass , and are non - spinning , is well characterized @xcite .
there have also been successes modelling the merger of neutron star binaries and neutron star black hole binaries @xcite .
numerical investigations of the full parameter space of compact binary coalescence waveforms are ongoing .
as described above , the gravitational waveform for coalescing binaries is well modelled analytically and numerically . signal processing theory @xcite advocates the use of matched filtering to extract known signals from gaussian noise . matched filtering provides the backbone of searches for coalescing binaries @xcite .
however , two substantial challenges remain : searching over the large parameter space of coalescing binary signals , and dealing with non - stationarities in the data .
the full binary coalescence waveform depends upon as many as seventeen parameters and it remains a challenge to search efficiently over the full parameter space . while some parameters , such as the amplitude and coalescence phase of the waveform , can be extracted using analytical techniques , others can only be searched by repeatedly evaluating the matched filter at numerous points across the parameter space .
this is facilitated by creating a bank of template waveforms , subject to the condition that for any candidate signal only a small fraction of the signal ( typically 3% ) is lost due to filtering with a mis - matched waveform @xcite .
this method works well for binaries with non - spinning components .
however , the parameter space of spinning binaries is considerably larger .
several methods of attacking this problem have been proposed , including phenomenological waveforms @xcite , restrictions to binaries with a single spin @xcite and markov chain monte carlo @xcite techniques .
however , the increase in both computational cost and the background rate associated with covering the spin parameter space has , to date , rendered this unfeasible .
indeed , it was been shown that searching for spinning binaries with non - spinning waveform templates provides comparable sensitivity to the currently available spinning searches the benefits of using the improved , spinning template model are negated by the increase in false alarms , particularly in real data @xcite .
investigations of the use of spinning templates in gravitational wave searches continue .
the data from gravitational wave interferometric detectors contains a significant number of non - stationary transients caused by various environmental and instrumental sources .
these reduce the sensitivity of a matched filter search as loud noise transients will produce a large signal to noise ratio , even if they do not match well the gravitational wave binary coalescence signal .
many techniques have been developed to has mitigate the effect of these noise transients in the data . among the most powerful
are : data quality and veto tests which flag times of poor data quality or use auxiliary channels with known couplings to the gravitational wave channel to remove times of poor data @xcite , as described in section [ ssec : burst_methods ] ; coincidence tests which require that a signal be observed , with consistent parameters , at widely separated sites @xcite ; signal consistency tests which compare the observed signal in the detector to the predicted waveform @xcite ; the use of improved ranking statistics which better separate the foreground and background by taking into account additional information , over and above the signal to noise ratio of the candidate @xcite . by making use of these additional tests , searches for binary coalescences are approaching the theoretically predicted sensitivity in gaussian noise .
gravitational wave data from the geo , ligo , tama and virgo interferometric detectors have been analyzed for coalescing binary signals . to date , no gravitational wave signal has been observed .
consequently an ever improving set of upper limits has been placed on the rate of binary coalescence as a function of the mass of the binary .
upper limits have been derived for systems ranging from neutron star binaries through to intermediate mass black hole binaries . here , we recap the latest results and compare them with astrophysical predictions .
astrophysical estimates of binary neutron star coalescence rates can be derived from observations of binary pulsars in the galaxy .
these rates are extrapolated to the local universe under the assumption that the rate of binary coalescence follows the star formation rate in spiral galaxies , which is obtained from measurements of the blue light luminosity of galaxies @xcite .
thus , results are quoted per @xmath27 per year , where @xmath28 times the solar blue luminosity and , for reference , the milky way is approximately @xmath29 .
the predicted rate for binary neutron star coalescence is @xmath30 , although the rate could plausibly be as much as an order of magnitude larger @xcite . to date , there are no direct observations of black hole - neutron star or black hole - black hole binaries . thus , rate estimates are based upon population synthesis , and yield realistic rates of @xmath31 for neutron star - black hole binaries @xcite and @xmath32 for binary black holes @xcite . in both cases , the rate could feasibly be as high as @xmath33 .
the first eighteen months of the ligo s5 data have been analyzed for gravitation wave signals from coalescing binaries with a total mass less than @xmath34 @xcite .
upper limits obtained by combining the result of this analysis with results from s3 and s4 provide the most stringent bounds on the coalescence rate from gravitational wave observations .
the binary neutron star coalescence rate is restricted , at 90% confidence , to be less than @xmath35 @xcite .
this is a factor of thirty above optimistic rates , and several hundred above the best estimate of the rate .
it is , however , interesting to note that the upper limit has improved by four orders of magnitude from the one obtained with ligo s first science run @xcite . for binary black holes , with component masses @xmath36 , the 90% rate limit is @xmath37 , and for black hole - neutron star binaries , the limit is @xmath38 @xcite .
these limits are again between one and two orders of magnitude from the upper end of astrophysical predictions , and three orders of magnitude from best estimates . the most recent upper limits for binaries with a total mass greater than @xmath34 were obtained using data from the fourth ligo science run . for binaries of total mass between @xmath39 and @xmath40
a search with a phenomenological template family @xcite of binary black hole waveforms gave an upper limit of @xmath41 @xcite .
for higher masses , a search for the ringdown portion of the signal yielded a rate limit for binary coalescences in the mass range @xmath42 to @xmath43 of @xmath44 .
a search of the ligo s5 data for black hole binaries with a total mass up to @xmath26 is being pursued @xcite .
this search will , for the first time , make use of full inspiral - merger - ringdown coalescence waveforms obtained by enhancing the post - newtonian inspiral waveforms with merger and ringdowns simulated numerically @xcite .
the coalescence of two neutron stars or a neutron star and a black hole is one of the preferred progenitor scenarios for short - duration grbs @xcite . by making use of the known time and sky location of observed grbs
, it is possible to perform a more sensitive search of the gravitational wave data .
this has been done for grb 070201 , which was a short grb localized in a region of the sky which overlapped the andromeda galaxy @xcite .
the search yielded no evidence of gravitational waves , and allowed for the exclusion of a binary coalescence progenitor in m31 with 99% confidence .
recently data taking with the enhanced ligo and virgo+ detectors began .
it is hoped that these detectors will achieve a factor of two sensitivity improvement over the initial detectors , which translates to almost an order of magnitude increase in the volume of the universe that can be probed for binary coalescences . in the following years
, advanced gravitational wave detectors will bring an order of magnitude increase in sensitivity over the initial configurations . at this stage ,
astrophysical estimates predict the observation of gravitational waves from tens of binary coalescences per year .
although neutron stars in coalescing binary systems represent a relatively well - understood population of putative gravitational wave sources , with well - defined gravitational luminosities and population statistics , the same neutron stars ( and even isolated neutron stars ) can in principle radiate a detectable amount of gravitational radiation well before coalescence .
radio and x - ray pulsar populations give us only a hint of the vast number ( @xmath45 ) of neutron stars that exists in the galaxy .
we currently see perhaps only one in a million neutron stars as a pulsar , but any neutron star , pulsar or not , can generate continuous quasi - sinusoidal gravitational waves through rotation .
any non - axisymmetric spinning neutron star will generate gravitational radiation .
although the centrifugal deformation can be expected to make it significantly oblate ( @xmath46 ) , this axisymmetric deformation will not itself generate gravitational radiation .
instead one requires the shape of the neutron star to be supported against relaxation to a fluid equipotential surface by a force , possibly an elastic stress force from the crust of the star , a magnetic force distorting the crustal shape or possibly a distortion caused by accretion or gravitational radiation - driven instabilities .
a neutron star with its spin axis oriented towards us , at a distance @xmath47 , with such a mass quadrupole moment @xmath48 around its axis of spin will generate a circularly polarised gravitational signal , at a frequency equal to twice the rotation rate @xmath49 of the star , with amplitude @xmath50 if the spin axis is inclined to the line - of - sight by an angle @xmath51 , the radiation becomes elliptically polarised and the amplitude is reduced .
it is often convenient to express @xmath48 as the product of an axial moment of inertia , @xmath52 , and an effective equatorial ellipticity , @xmath53 where @xmath54 and @xmath55 are the other two principal moments of inertia .
there are clearly two very important questions here : i ) what is the size of deformation that neutron stars _
could _ have , given their equation of state , crystalline structure and physical environment , and ii ) what deformations do they _ actually _ have .
the answer to the first question depends critically on both the equation of state of the neutron star ( and whether it is indeed a neutron star or a quark star ) and its crystalline structure @xcite .
recent work by horowitz and kadau @xcite has indicated that the structure may indeed by highly crystalline , with point defects rapidly squeezed out to give breaking stains of as much as 0.1 .
this would allow self - supported equatorial ellipticities of perhaps @xmath56 to @xmath57 .
this second question is harder to address . unlike in binary neutron star systems such as psr b1913 + 16 , psr j0737@xmath583039 , psr b1534 + 12 and psr j1756@xmath582251 , where the orbital evolution presents convincing evidence of the very early stages of a coalescence
, we have no direct evidence of spin - gravitars ( that is , neutron stars whose observed spin - down is well - modelled by gravitational braking ) .
certainly in the case of equatorial deformation supported by crustal strength we can not dismiss the notion that some neutron stars have perfectly annealed equipotential surfaces , with negligible axial quadrupole moment and therefore essentially no gravitational luminosity .
for example , the extremely low period derivatives of millisecond pulsars hint that these neutron stars at least show very little equatorial asymmetry . at some level
all neutron stars will show deformation due to internal magnetic pressures , though this only becomes relevant for the strongest of magnetars @xcite .
however , the story is less clear for young and/or accreting neutron stars .
for both these classes there are plausible mechanisms to supply both the energy and the deformation necessary for significant gravitational luminosity ( see for example @xcite ) . for an isolated neutron star we may postulate that the gravitational luminosity should be less than the rate of loss of rotational kinetic energy for a rigid body . in turn
this defines an upper limit on the strain amplitude we could expect from rigid body gravitar with a spin - down rate @xmath59 as @xmath60 this `` spin - down upper limit '' falls below the 1-year strain sensitivity of both the initial ligo and virgo detectors for all but a small number of known radio pulsars , making these pulsars unlikely first - detection candidates .
however , only a tiny fraction of the neutron star population is seen as electromagnetic pulsars , leaving the possibility of a gravitationally luminous , but electromagnetically dim , population .
there is an obvious sense in which it is easier to search for a continuous quasi - sinusoidal signal than a transient inspiral or burst signal .
a long - lived signal can be re - observed ( or even retrospectively observed using archived data ) and confirmed as astrophysical .
in addition , there are radio and x - ray pulsars that accurately trace the rotational evolution of around 100 pulsars whose gravitational signal would fall in terrestrial observing bands .
however , the benefits of a continuous wave search stop there .
an inevitable consequence of searching for a long - duration deterministic signal containing up to @xmath61 cycles is an exquisite sensitivity to its parameter values .
most obviously , a change in frequency corresponding to just one more or one fewer cycles during the observation would represent an entirely different search , with a template for the expected signal that , as a matched filter , was insensitive to the original signal .
it also becomes apparent that if the neutron star is spinning down , the number of spin - down templates necessary to cover the possible alternatives scales as the square of the observing time .
additionally , the doppler modulation of the received signal is sensitive to the position of the source on the sky . for year - long observations
this angular sensitivity is approximately the gravitational diffraction limit of an aperture the diameter of the earth s orbit about the sun ( @xmath62arcsec at typical frequencies ) . if we do not have the benefit of a radio trace of the neutron star s rotational evolution and sky location we are forced to perform a search over this parameter space , and it rapidly becomes apparent that the parameter space is huge .
continuous - wave searches are by far the most computationally expensive searches that the gravitational wave community undertakes , and in its most general form the problem is ( and always will be ) fully limited by available computing power .
no matter what its form , any coherent search method will improve its strain signal - to - noise ratio as @xmath63 where @xmath64 is the detector s ( strain ) power spectral density at the frequency of the signal and @xmath10 is the observing time .
however the overall sensitivity of a search is not solely dependent on signal - to - noise ratio .
the more trials that are undertaken ( i.e. , templates that are searched ) the greater the probability of random noise popping up to unluckily appear like a signal .
the apparent signal - to - noise ratio indicative of a true signal is therefore several tens for searches that pick the strongest candidate from over a wide parameter space .
one consequence of this is that any convincing signal must have a relatively high signal - to - noise ratio after only a relatively short coherent integration .
this allows one to combine these short integrations incoherently ( as powers , ignoring phase ) without too great an impact in overall sensitivity and develop semi - coherent search methods which are computationally much cheaper @xcite .
the overall sensitivity does however only improve as the quarter power of the number of incoherently combined contributions .
ground - based cw search efforts have concentrated on variants of the above , from fully coherent long - timescale searches for gravitational wave signals phase - locked to radio pulsars @xcite to searches concentrating on non - pulsing targets @xcite and massively computational all - sky searches using a variety of semi - coherent techniques @xcite .
upper limits on the strength of continuous gravitational waves from both known and unknown galactic neutron stars have been made regularly since the first ligo / geo science run in 2002 . as sensitivities and run lengths have improved , the limits have steadily dropped .
the recent 23-month ligo s5 run had sufficient sensitivity to show that the crab pulsar is not a gravitar ( i.e. , is not spinning down solely due to the emission of gravitational radiation ) . in itself
this is no surprise the overall energy budget of the crab nebula and pulsar has to account for the nebula luminosity and expansion .
however , the early s5 result ( covering just the first 9 months of data ) was sufficiently sensitive to show that less than about 3 percent of the spin - down luminosity of the crab pulsar is due to gravitational emission @xcite .
the full s5 result , with a fully coherent search for gravitational emission from 116 known pulsars , including the crab pulsar , is expected soon .
additional work is going on to use these and other newly - developed targeted algorithms to search virgo vsr1 data for emission from the vela pulsar at @xmath65hz @xcite .
the s5 run has also resulted in the most sensitive `` all sky ''
( i.e. , survey ) searches to date .
two early - s5 papers have already been published on this @xcite .
the first comprised a semi - coherent search , incoherently adding 30-minute demodulated power spectra ( the `` power flux method '' ) .
the second was an early result from einstein@home , a distributed screen saver application that currently attracts about 200000 users worldwide and returns @xmath66tflops to search project .
this method looked for coincident detections between multiple 30-hour coherent searches .
both these all - sky searches returned strain upper limits of around @xmath67 for a wide spectral band , and these are levels with real astrophysical significance . a simple argument , originally by blanford but developed by knipsel and allen @xcite indicates that for our galaxy population and distribution of neutron stars , the loudest expected cw source would have a strain at earth at about this level ( under certain assumptions ) .
such a source would need to be within a few hundred parsecs of earth .
in addition to targeted and all - sky searches , more specialised `` directed '' search methods are also being used to tackle likely sky locations including globular clusters , the low - mass x - ray binary sco - x1 , the galactic centre and supernova remnants .
one such search , for gravitational emission from the x - ray point source at the centre of cas a is nearing completion and has reported an expected sensitivity also in the range @xmath68 @xcite .
the ligo and virgo collaborations have now developed a broad suite of algorithms and methods to tack a wide range of potential sources of continuous gravitational radiation , including all - sky searches for binary sources , and the full power of these will be applied to data form the current s6/vsr2 runs . as with other searches that involve population statistics
, the crude extrapolation holds that a factor @xmath69 improvement in sensitivity will increase detection numbers by a factor @xmath70 .
clearly the physical extent of the galaxy places an upper limit on this , but that only becomes relevant for current all - sky searches when broadband sensitivities are a factor @xmath66 times their current values .
perhaps more important is a consideration of the types of neutron star that may be detectable in the future using instruments with an improved low frequency response .
current detectors show good sensitivity only to relatively rapidly spinning pulsars , most of which are recycled millisecond pulsars with low observed spin - down rates and , probably , low gravitational luminosity .
young , glitchy pulsars are more common at gravitational frequencies below @xmath66hz , with some of the most interesting , rapidly braked , sources closer to 10hz , so the low - frequency wall is a particular challenge for future continuous wave gravitational observations .
a stochastic gravitational - wave background ( sgwb ) refers to a long - lived random gw signal .
this is generally produced by a superposition of many unresolved sources , and can be characterized as cosmological or astrophysical according to the epoch in which the gws are generated .
cosmological backgrounds can be assumed to be approximately isotropic , unpolarized and stationary , while astrophysical backgrounds may have additional structure depending on the nature of their sources .
a convenient measure of the strength of a sgwb is the energy density in the gws , per logarithmic frequency interval , in units of the critical energy density needed to close the universe : depends on the value of the hubble constant ; it has become conventional to use the fiducial value @xmath71 when defining @xmath72 . ]
@xmath73 cosmological models which produce a sgwb include amplification of quantum vacuum fluctuations during inflation @xcite , phase transitions @xcite , pre - big - bang models @xcite , and cosmic ( super-)string models @xcite .
standard inflationary models generate a backround of constant @xmath72 over many decades of frequencies , but the amplitude of such a background is already bounded by cosmic microwave background observations to be @xmath74 @xcite .
astrophysical gw backgrounds can be generated by unresolved superpositions of sources such as cosmic string cusps @xcite , supernovae @xcite , and neutron - star instabilities @xcite .
the most stringent indirect limit on a sgwb in the frequency range of ground - based detectors comes from a constraint on the total energy density present at the time of nucleosyntheis .
this big - bang nucleosynthesis ( bbn ) bound limits the total energy density in gravitational waves to be @xmath75 where the effective number @xmath76 of neutrino species at bbn is constrained to by @xmath77.@xcite note that this limit only applies to cosmological sgwbs , i.e. , gravitational waves generated before the era of nucleosynthesis .
since the amplitude of a sgwb will be much smaller than that of instrumental noise in a typical ground - based detector , one needs to exploit the expectation that while instrumental noise will be ( predominantly ) uncorrelated between independent detectors , the gravitational wave signals in a pair of detectors should have an average correlation @xmath78 where @xmath79 encodes the observing geometry ( location and orientation of detectors 1 and 2 , and in the case of an anisotropic background , the spatial distribution of the background ) and @xmath80 is a one - sided power spectral density for the sgwb which is given for an isotropic background by @xmath81f^{-3}{\omega_{{\hbox{\scriptsize\rm gw}}}}(f ) \ .\ ] ] the standard search method @xcite for an isotropic background cross - correlates the data from pairs of detectors using an optimal filter @xmath82 where @xmath83 are the noise power spectra for the two detectors and @xmath84 is the expected shape of the sgwb spectrum .
the resulting search is sensitive to a background @xmath85 of strength @xmath86 ^ 2}{s_1(f)\,s_2(f ) } \right)^{-1/2 } \ .\ ] ] note that the sensitivity of a cross - correlation search improves like the square root of the observing time @xmath10 .
also , stochastic background measurements tend to be dominated by the low end of the available frequency range , because @xmath79 oscillates with increasing @xmath87 within an envelope whose leading term is @xmath88 and because eq . [ sgw ] means that a constant-@xmath72 background has @xmath89 .
a cross - correlation search can also be used to search for an astrophysical background with a specified spatial distribution , e.g. , a sgwb coming from one point on the sky @xcite .
more sophisticated techniques can be used to recover the spatial distribution of a measured background @xcite .
the most stringent direct limit on @xmath72 was set using data from the s5 run of ligo livingston and ligo hanford @xcite , which set the 95% confidence level upper limit of @xmath90 assuming @xmath72 to be constant over the interval @xmath91 .
a sgwb of the excluded strength , confined to those frequencies , would contribute @xmath92 to the total value of @xmath93 .
this limit is therefore more stringent than the bbn bound ( eq . [ e : bbnbound ] ) and we have entered the era where ground - based gw detectors are placing new limits on gravitational wave backgrounds of cosmological origin . the previous limit from s4 data @xcite , while less stringent by about an order of magnitude , already placed new restrictions on the parameters of some cosmic string models which generate gws both before and after the era of nucleosynthesis . additional searches of s4 ligo data set limits on the strength of possible point - like backgrounds @xcite and ( by correlating ligo livingston data with data from the allegro bar detector ) set a higher - frequency limit of @xmath94 @xcite .
correlation measurements using ligo and virgo data are expected to improve the high - frequency measurement @xcite .
further searches for anisotropic backgrounds are also being conducted .
the current search for gw covers multiple types of signals originating from different possible astrophysical events like core collapse of massive stars and neutron star formation , coalescing binary systems of neutron stars and black holes , non - axysimmetric spinning neutron stars and signals produced by a large collection of incoherent sources .
the data acquired by the most sensible gw observatories , which are at present the ligo and virgo interferometers , are analysed applying different methods and strategies targeted to the identification and characterisation of the signals emitted by these possible sources .
moreover , methods able to catch signals coming from unknown sources are also currently used .
the analysis of the latest scientific data , acquired during the first part of the s5/vsr1 run , did not provide any evidence of a possible detection .
upper limits on the rate of events and/or the strain amplitude @xmath95 are then derived and compared to the astrophysical predictions .
these limits are already valuable scientific results which reinforce and widen our knowledge of the astrophysical events involved .
for example they are now approaching the plausible astrophysical values in the case of gw originating from binary coalescing systems and have already set a bound to the percentage of the spin - down luminosity of the crab pulsar on the energy emitted in gravitational radiation . the energy density in a stochastic gw background
around 100 @xmath0 has been constrained to a limit which is more stringent than the the big - bang nucleosynthesis bound , the strongest indirect limit at those frequencies . from the analysis of the data in coincidence with the grb 070201 it has been possible to exclude the hypothesis of a binary merger in m31 as the progenitor of this event .
further results from the analysis of the full s5/vsr1 run data are expected soon . meanwhile , the analysis of the data from the s6/vsr2 run will probably , if no gw detections are found , improve the current limits , owing to the possible improvement of sensitivity of ligo and virgo detectors .
the sensitivity of the present and future gw detectors gives the possibility of studying astrophysical events jointly with other observatories , i.e. electromagnetic and neutrino observatories , likely bringing additional information on the physics of the sources and on their characteristics .
the future searches will then open the possibility of performing a mature gw astronomy .
the authors would like to thank their colleagues in the ligo and virgo scientific collaborations , especially ray frey , ben owen and joe romano for comments and suggestions .
sf would like to acknowledge the support of the royal society .
jtw is supported by nsf grant phy-0855494 and by the college of science of rochester institute of technology .
103 abbott , b.p . , et al .
phys . * 72 * , 076901 ( 2009 ) acernese , f. , et al . : class .
* 23 * , s635 ( 2006 ) takahashi .
r. and the tama collaboration , class .
* 21 * , s403 ( 2004 ) lck , h. , et al . : class .
* 23 * , s71 ( 2006 ) ott , c.d .
, et al . :
lett . * 96 * , 201102 ( 2006 ) ott , c.d . , et al .
: phys . rev .
lett . * 98 * , 261101 ( 2007 ) ott , c.d . , et al .
: class .
* 24 * , 139 ( 2007 ) marek , a. , janka , h .-
, mller , e. : astron .
astrophys . * 496 * , 475 ( 2009 ) baker , j. g. , campanelli , m. , pretorius , f. , zlochower , y. : class .
* 24 * , s25 ( 2007 ) pretorius , f. in m. colpi , p. casella , v. gorini , u. moschella , and a. possenti , editors , _ physics of relativistic objects in compact binaries : from birth to coalescence_. springer , heidelberg , germany , 2009 .
baiotti , l. , giacomazzo , b. , rezzolla , l. : phys .
d * 78 * , 084033 ( 2008 ) klimenko , s. , mitselmakher , g. : class . quant .
* 21 * , s1819 ( 2004 ) klimenko , s. , et al . : class .
grav . * 25 * , s114029 ( 2008 ) chatterji , s. , et al . : class .
* 21 * , s1809 ( 2004 ) clapson , a .- c . , et al . : class .
25 * , 035002 ( 2008 ) chatterji , s. , et al .
d * 74 * , 082005 ( 2006 ) cavalier , f. , et al .
d * 74 * , 082004 ( 2006 ) markowitz , j. , zanolin , m. , cadonati , l. , katsavounidis , e. : phys .
d * 78 * , 122003 ( 2008 ) acernese , f. , et al .
: class .
* 26 * , 085009 ( 2009 ) brady , p.r . , creighton , j.d.e . , and wiseman , a.g .
: class .
* 21 * , s1775-s1782 ( 2004 ) abbott , b.p . , et al .
: arxiv:0905.0020 [ gr - qc ] .
abbott , b.p . , et al .
: arxiv:0904.4910 [ gr - qc ] .
abbott , b.p . , et al .
: astrophys .
j. * 681 * , 1419 ( 2008 ) abbott , b.p . , et al .
: phys . rev .
d * 77 * , 062004 ( 2008 ) acernese , f. , et al .
: class .
quant . grav .
* 25 * , 225001 ( 2008 ) abbott , b.p . , et al .
* 101 * , 211102 ( 2008 ) abbott , b.p . , et al .
: astrophys .
j. * 701 * , l68 ( 2009 ) abbott , b.p . , et al .
: arxiv:0904.4718v2 [ astro-ph.co ] .
ushomirsky , g. , cutler , c. , bildsten , l. : mon . not .
r. astron .
* 319 * , 902 ( 2000 ) owen , b.j .
95 * , 211101 ( 2005 ) horowitz , c.j . ,
kadau , k. : phys .
* 102 * , 191102 ( 2009 ) colaiuda , a. , ferrari , v. , gualtieri , l. , pons , j.a .
: mon . not .
r. astron .
soc . * 385 * , 2080 ( 2008 ) friedman , j.l . , schutz , b.f .
: astrophys .
j. * 222 * , 281 ( 1978 ) bildsten , l. : astrophys .
j. lett .
* 501 * , l89 ( 1998 ) owen , b.j .
: class .
* 23 * , s1 ( 2006 ) cutler , c. , gholami , i. , krishnan , b. : phys .
d * 72 * , 042004 ( 2005 ) abbott , b.p .
, et al . :
d * 69 * , 082004 ( 2004 ) abbott , b.p .
, et al . :
* 94 * , 181103 ( 2005 ) abbott , b.p . , et al .
: phys . rev .
d * 76 * , 042001 ( 2007 ) abbott , b.p .
, et al . :
. j. lett .
* 683 * , l45 ( 2008 ) abbott , b.p . , et al .
d * 76 * , 082001 ( 2007 ) abbott , b.p .
, et al . :
d * 76 * , 082003 ( 2007 ) abbott , b.p .
, et al . :
d * 72 * , 102004 ( 2005 ) abbott , b.p .
, et al . :
d * 77 * , 022001 ( 2008 ) abbott , b.p .
, et al . :
d * 79 * , 022001 ( 2009 ) abbott , b.p .
lett . * 102 * , 111102 ( 2009 ) abbott , b.p . , et al .
d * 80 * , 042003 ( 2009 ) frasca , s. , for the ligo scientific collaboration and virgo collaboration : ligo - g0900712-v1 ( 2009 ) + http://dcc.ligo.org/cgi-bin/docdb/showdocument?docid=g0900712 knispel , b. , allen , b. , phys .
d * 78 * , 044031 ( 2008 ) wette , k. , et al . : class .
* 25 * , 235011 ( 2008 ) allen , b. and romano , j. , phys .
d * 59 * , 102001 ( 1999 ) ballmer , s. , class .
* 23 * , s179 ( 2006 ) mitra , s. , et al .
d * 77 * , 042002 ( 2008 ) abbott , b.p .
, et al . :
nature * 460 * , ? ? ? - ? ? ? ( 2009 ) abbott , b. , et al .
: astrophys .
j. * 659 * , 918 ( 2007 ) abbott , b. , et al .
d * 76 * , 082003 ( 2007 ) abbott , b. , et al .
d * 76 * , 022001 ( 2007 ) cella , g. et al . : class .
* 24 * , s639 ( 2007 ) | the first generation of gravitational wave interferometric detectors has taken data at , or close to , their design sensitivity .
this data has been searched for a broad range of gravitational wave signatures .
an overview of gravitational wave search methods and results are presented .
searches for gravitational waves from unmodelled burst sources , compact binary coalescences , continuous wave sources and stochastic backgrounds are discussed . |
the quantum circuit model is a standard model of quantum computation describing the relationship between input and output by a sequence of elementary gates .
this model is widely used since it is shown to be universal , and it has a good correspondence to the logic circuit of classical computation .
however at the same time , there is a restriction that elementary gates have to be performed from left to right without creating any loops in the quantum circuit in the quantum circuit .
thus only operations with definite causal order can be performed , whereas such a restriction may not be necessary in quantum mechanics as pointed out in @xcite .
the role of causal order in quantum mechanics has been widely studied , since the existence of time - loops is allowed by the theory of general relativity .
historically , this question was motivated by deutsch in @xcite , where he considered quantum computation with closed timelike curves ( ctcs ) .
bennet , schumacher , and svetlichny considered a different model of ctcs @xcite , referred to as bss ctcs @xcite .
others also refer the same model as p - ctcs ( closed timelike curves via quantum postselection ) @xcite , where a classical time - paradox can be resolved by considering quantum computation with ctcs .
the computational power of ctcs was studied in @xcite , where a quantum circuit with deutsch - type ctcs is shown to be able to solve a computational problem belonging to the pspace class in polynomial time .
quantum computation augmented with postselection is shown to have a power to to solve problems in the pp class in polynomial time @xcite . even besides ctcs
, we can consider a situation with strange causality .
for example , quantum communication without causal structure is formulated in @xcite . in @xcite , an operation beyond causally ordered quantum computation
called quantum switch has been investigated .
quantum switch is a super - map of which input is two different single - qubit unitary gates @xmath2 and @xmath3 and output is a two - qubit controlled - unitary that coherently performs
two differently ordered unitary operations @xmath1 and @xmath0 depending on the state of a control - qubit .
it was proven that quantum switch can not be implemented within the quantum circuit model with a fixed causal order , if each of @xmath2 and @xmath3 is allowed to be used only once .
are there any operations that can not be implemented within the causally restricted quantum circuit model but still possible in quantum mechanics ? in this paper , we show a positive result that by implementing the quantum computation adiabatically one can map the causal order of operations to a construction of hamiltonian , and that the causal order can be manipulated to implement quantum switch introduced in @xcite .
to achieve this task , we analyze adiabatic gate teleportation ( agt ) , a model of quantum computation proposed by bacon and flammia @xcite . in agt ,
a single - qubit gate is applied to an unknown quantum state of an input qubit and transferred to an output qubit via an intermediate qubit by adiabatically changing interactions between qubits from a fixed initial hamiltonian to a final hamiltonian determined by the gate . based on agt ,
we develop parallelized adiabatic gate teleportation ( pagt ) , a scheme where consecutive gate teleportations are implemented in one adiabatic shift from the initial hamiltonian to the final hamiltonian by introducing a parallelizable _ gate hamiltonian _ implementing a single - qubit gate . in pagt ,
causal order of gate operations are mapped to spatial order of interactions in the final hamiltonian .
all gate hamiltonians are simultaneously applied , although the speed of adiabatic evolution should be slowed down due to decreasing energy gaps .
thus pagt does not contribute to speed up performing a sequence of gate operations by parallelizing , but to eliminate the control of causal order of gate operations in the time domain . using this property
, we also present the controlled - pagt scheme that performs the controlled - unitary operations implemented by quantum switch . as a special case of controlled - pagt
, we show that a controlled - unitary gate is implementable for an unknown unitary gate if the unitary gate is guaranteed to be written in a orthogonal matrix form .
this paper is organized as follows .
we present preliminaries to review the matrix representation of bipartite states , the adiabatic theorem , and adiabatic gate teleportation in section [ preliminaries ] . in section [ sec : gateham ]
, we introduce a parallelizable gate hamiltonian and analyze its property . in section [ pagt ] and section [ cpagt ]
, we present our main results , the pagt scheme and the controlled - pagt scheme , respectively .
we show several examples of the controlled - pagt scheme in section [ applications ] .
we summarize the results in section [ conclusion ] .
we use the matrix representation of bipartite pure states @xcite for simplifying descriptions of our scheme . consider a bipartite system represented by a hilbert space given by @xmath4 where @xmath5={\rm dim}[\mathcal{h}_2]=d$ ] and
let a , b and c be @xmath6 matrices . by using a set of coefficients satisfying @xmath7
, we denote a bipartite pure state in @xmath4 in orthonormal bases @xmath8 for @xmath9 and @xmath10 for @xmath11 by @xmath12 the matrix @xmath13 determines a bipartite state and can be considered as a linear map from @xmath11 to @xmath9 .
note that in our definition eq . , the normalization factor @xmath14 has been introduced to guarantee the state represented by @xmath15 to be normalized for simplifying descriptions for @xmath5={\rm dim}[\mathcal{h}_2]=d$ ] , whereas in the original definition of the matrix representation of bipartite states presented in @xcite , @xmath15 is not necessary to be normalized but applicable for more general cases of @xmath5 \neq { \rm dim}[\mathcal{h}_2]$ ] .
this representation of bipartite states has a convenient property @xmath16 where @xmath17 , @xmath18 , and @xmath19 denotes the transposition of a matrix . for a maximally
entangled state @xmath20 where @xmath21 represents the identity matrix , @xmath22 where @xmath23 represents the identity matrix on @xmath24 , and @xmath25 where @xmath26 is an arbitrary unitary operation hold .
we note that the matrix representation is basis dependent since the description of a matrix depends on the choice of the bases @xmath27 and @xmath28 . in the followings
, we mainly consider qubit systems ( @xmath29 ) and take the computational bases of qubits for @xmath9 and @xmath11 .
adiabatic quantum computation was first proposed in @xcite , as a model to solve the sat problem by using adiabatic evolution .
the evolution is governed by time - dependent hamiltonian represented by @xmath30 , which starts from the initial hamiltonian @xmath31 and ends with the final hamiltonian @xmath32 .
the parameter @xmath33 is a function of time @xmath34 , varying from 0 to 1 with total time @xmath19 , i.e. @xmath35 and @xmath36 .
the input state is the ground state of @xmath31 , which is assumed to be easily prepared , and the solution of the problem is encoded in the ground state of @xmath32 .
the main assumption of this algorithm is that the adiabatic theorem is valid during evolution . roughly speaking , the adiabatic theorem states that ground state of time - dependent hamiltonian remains its instantaneous ground state at later time provided that the change of the hamiltonian is _
slow enough_. however , necessary and sufficient conditions for this slowness are not yet fully understood . a traditional quantitative condition presented in @xcite for the adiabatic theorem is @xmath37 } |\langle e_1(s)|\frac{dh(s)}{ds } |e_0(s)\rangle}{{\mathcal{g}}^2 } \right ) , \end{aligned}\ ] ] where @xmath38 is the instantaneous @xmath39-th energy eigenstate ( @xmath40 represents the ground state ) and @xmath41 } \ { \delta(h(s ) ) \}=\min_{s\in[0,1 ] } ( e_1(s)-e_0(s))$ ] is the minimum energy gap between the ground state and the first excited state during the evolution .
however , it has been pointed out that this quantitative condition fails in certain situations @xcite .
recently , the necessity of this condition has been considered along with new sufficient conditions @xcite . in @xcite ,
the three conditions given by @xmath42 , { \notag \\}&(b ) & & \int_0^t \left| \left ( \frac{\langle{e_n(t)}|{\dot{e}_m(t)}\rangle}{e_n(t)-e_m(t ) } \right)'\right| dt<\!\ !
< 1 , { \notag \\}&(c ) & & \int_0^t \left|\frac{\langle{e_n(t)}|{\dot{e}_m(t)}\rangle}{e_n(t)-e_m(t)}\right| \left| \langle{e_m(t)}|{\dot{e}_l(t)}\rangle \right|dt<\!\ !
< 1 { \notag \\}&&&m \neq n , m \neq l\end{aligned}\ ] ] have been shown to be sufficient and ( a ) has been shown to be necessary @xcite .
these functions are hard to calculate in general , since it is difficult to compute the instantaneous energy eigenstate . in this paper
, we refer to a stronger condition for the adiabatic theorem presented in @xcite .
consider the time - dependent hamiltonian @xmath43 , where @xmath44 has a unique ground state @xmath45 $ ] .
then for any fixed @xmath46 , if @xmath47 is satisfied , the final state under the adiabatic evolution at the time @xmath19 is @xmath48-close in @xmath49-norm to the ground state of @xmath32 .
the matrix norm @xmath50 denotes the spectral norm defined by @xmath51 .
the above conditions are derived for a linear interpolation function @xmath52 .
the function @xmath53 can be chosen in a smarter way so that it varies faster when the energy gap is large and slower when the energy gap is small .
when we change the speed of @xmath53 depending on the energy gap , the total time sufficient for the adiabatic theorem to hold is evaluated by @xmath54 it was pointed out that the necessary time @xmath19 becomes proportional to @xmath55 , and not @xmath56 in a certain case . by adjusting @xmath53 on each infinitesimal time interval , the total calculation time of the adiabatic implementation of grover s search algorithm @xcite reduced from @xmath57 to @xmath58 , which is the same as the standard implementation of grover s algorithm using the quantum circuit model . in this section ,
we review adiabatic gate teleportation ( agt ) @xcite .
we consider a system consisting of only qubits in this paper , but one can easily generalize the scheme to general qudit systems .
we use @xmath59 and @xmath60 to represent the pauli operators , generators of su(2 ) .
subscripts are used to specify the hilbert space of states or operators . in case
it is not confusable , we often omit the identity operator and the symbol @xmath61 to represent the tensor product of operators .
for example , when we are considering a hilbert space @xmath62 , @xmath63 denotes @xmath64 note that @xmath63 does not represent a simple multiplication of two operators on the same hilbert space since the different subscripts imply that two operators act on two different hilbert spaces . before describing agt
, we introduce a special version of agt , adiabatic teleportation .
consider a three qubit system @xmath62 .
adiabatic teleportation aims to transfer unknown quantum state encoded in the first qubit @xmath65 ( input state ) to the third qubit ( output state ) using adiabatic evolution under a time - dependent hamiltonian of interacting qubits .
the second qubit acts as a mediator for transferring the state .
the initial state is prepared in @xmath66 , which is a ground state of the initial hamiltonian @xmath67 where @xmath68 .
note that we only consider the case @xmath69 .
this hamiltonian is free on @xmath9 , so it has two degenerate ground states .
we can choose an arbitrary input state @xmath70 , but the state of the second and the third qubits is fixed to be in @xmath71 . the initial hamiltonian given by eq .
( [ inithamil ] ) is a slightly modified from the one introduced in the original paper of agt @xcite given by @xmath72 .
the additional terms @xmath73 and @xmath74 in eq .
( [ inithamil ] ) just shift the energy eigenvalues of the initial hamiltonian by a constant , and do not change the eigenstates .
however , the existence of these additional terms is crucial for parallelizing agt .
we adiabatically change the initial hamiltonian to the final hamiltonian given by @xmath75 to connect the initial and final hamiltonian , we may introduce a time - dependent hamiltonian in the form of @xmath76 where a parameter denoted by @xmath33 is a function of time @xmath34 satisfying @xmath77 and @xmath78 with total time @xmath19 of the adiabatic evolution .
the final hamiltonian also has two degenerate ground states .
the degenerate ground states have kept during the evolution .
this degeneracy is kept during the evolution . by adiabatically changing @xmath33
, the final state turns out to be @xmath79 .
that is , an unknown state @xmath80 is transferred from @xmath9 to @xmath81 .
thus this scheme is called to be adiabatic teleportation . to show how this adiabatic teleportation scheme works
, we first check that the energy gap is finite throughout the evolution .
the energy gap between the ( degenerate ) ground states and the first excited state of hamiltonian eq .
is @xmath82 it takes minimum value @xmath83 at @xmath84 , so we always have a nonzero energy gap . according to the adiabatic theorem ,
the system remains in the ground states throughout the evolution if parameter @xmath33 is varied slow enough .
we have introduced the adiabatic theorem for a hamiltonian @xmath44 with a unique ground state , but @xmath85 has two degenerate ground states throughout the evolution .
it is necessary to guarantee that no transition between there two degenerate ground states happens during adiabatic evolution under @xmath44 for an unknown state @xmath80 to be faithfully transferred in adiabatic teleportation . to show this
, we use stabilizer formalism and define logical operators @xmath86 and @xmath87 .
the logical space spanned by the logical operators is preserved during the adiabatic scheme due to @xmath88=0 $ ] ( @xmath89 ) .
we first explain that these logical operators encode information of the unknown input state @xmath80 . in the initial stage @xmath90 ,
the ground state of @xmath31 is @xmath66 , which is stabilized by @xmath91 and @xmath92 .
namely @xmath93 which means that @xmath91 takes a value 1 at @xmath90 , so @xmath94 indeed represents @xmath95 , and the same logic works for showing that @xmath96 represents @xmath97 at @xmath90 .
thus logical operators encode information of the unknown state @xmath80 .
we note that the term @xmath98 in the initial hamiltonian consists of terms @xmath99 and @xmath74 .
when considering of a ground state of these terms , it is sufficient to stabilize only @xmath100 and @xmath92 , since other term , @xmath101 and @xmath102 , can be derived from these operators .
we next show that the encoded information is faithfully transferred to the qubit on @xmath81 at the end of the adiabatic evolution . in the final stage @xmath103 , hamiltonian of the system
is the final hamiltonian @xmath32 .
when we drag the hamiltonian @xmath44 slow enough so that the adiabatic theorem holds , the ground state is kept throughout the evolution and finally we get @xmath104 for the ground state of @xmath32 .
since the state @xmath105 is stabilized by @xmath106 and @xmath107 , we have @xmath108 x_3 = x_3 $ ] and @xmath109 z_3 = z_3 $ ] at @xmath103 .
therefore , the information @xmath95 and @xmath97 of the input state on @xmath9 is sent to @xmath110 and @xmath111 on @xmath81 respectively through the evolution , which concludes the proof . here
, we describe the relationship between adiabatic teleportation and usual quantum teleportation.to start with , we explain usual quantum teleportation scheme .
[ fig : teleportation ] is the quantum circuit representation of quantum teleportation scheme .
alice wants to send her unknown state @xmath80 to bob ( receiver ) .
alice shares a first qubit which is unknown state @xmath112 , and a second qubit which is a part of maximally entangled state @xmath71 .
bob shares a third qubit which is the other side of a maximally entangled state .
next , alice measures her two qubits by bell basis @xmath113 which are described by , @xmath114 and @xmath115 she get a measurement outcome @xmath116 ( @xmath117 is either @xmath118 or @xmath60 ) with probability 25% for each , then bob s state collapses to @xmath119 . at last , bob performs a correction @xmath117 on his third qubit according to alice s measurement outcome to recover the unknown state @xmath80 .
we explained usual quantum teleportation scheme above .
however , we restrict our attention in the case when alice measures the state @xmath104 , bob gets the state @xmath120 without any correction . this effect is considered as p - ctcs ( closed timelike curves via quantum postselection ) @xcite , describing a time travel effect of a state @xmath80 from the final stage ( when the measurement is performed ) of the first qubit to the initial stage ( when the state @xmath71 is prepared ) of the third qubit via the second qubit mediator .
adiabatic teleportation is adiabatically mimicing this p - ctcs effect ( quantum teleportation and postselection ) by simultaneously applying hamiltonian on the whole system to drag the initial state @xmath66 to the final state @xmath121 .
while postselection can not be implemented deterministically in quantum circuit model , adiabatic evolution can mimic some class of quantum computation with definite measurement outcome by dragging the system with sufficiently long time determined by the energy gap .
represents alice s measurement outcome , which is either @xmath122 or @xmath123 .
bob s third qubit collapses by her measurement , here we can effectively treat bob s third qubit as being a state @xmath124 .
quantum teleportation of an unknown state @xmath80 from the first qubit to the third qubit is achieved after bob s correction @xmath117 .
note that in the case alice s measurement outcome is @xmath125 , bob does not need any correction to get a state @xmath80 on his third qubit . ]
we move on to explain agt scheme , which is obtained by slightly modifying adiabatic teleportation .
agt aims to obtain the output state @xmath126 where @xmath26 denotes a single - qubit gate ( unitary operation ) for an input state @xmath112 by adiabatic evolution .
the time - dependent hamiltonian for agt is given by the unitary conjugation form of eq . in terms of @xmath26 on @xmath81 , namely , @xmath127 the conjugation only affects the initial hamiltonian @xmath128 this modification is just a unitary conjugation changing only energy eigenstates and it keeps the energy eigenvalues .
thus the energy gap between the ground states and the first excited states is unchanged .
the correspondence of states between adiabatic teleportation and agt is @xmath129 at @xmath90 and @xmath130 at @xmath103 .
agt scheme adiabatically simulates the quantum circuit presented in fig .
[ agtcircuit](a ) , which is equivalent to the circuit in fig .
[ agtcircuit](b ) .
note that in fig .
[ agtcircuit](b ) , gate teleportation is valid applying @xmath26 on the third qubit even _ before
_ preparing an unknown state @xmath80 on the first qubit .
thus we can recognize that the circuit effectively achieves time travel of a state @xmath80 .
the circuit consists of teleporting a state @xmath80 from system @xmath9 to @xmath81 followed by a unitary operation @xmath26 applied to the teleported state . ]
when one considers solving some problems by quantum computation , there are some approaches to get solutions .
we review and clarify the difference between quantum circuit approach and adiabatic gate teleportation .
quantum circuit model tries to implement the unitary operation which solves the problem by describing a sequence of the elementary gates .
it describes input - output relation , while in adiabatic quantum computation model , we have a solution as a ground state of final hamiltonian .
we try to get this solution starting from the ground state of initial hamiltonian which is easy to prepare . computational concept of agt is more similar to the quantum circuit model than adiabatic quantum computation , because it implements a `` unitary gate '' , rather than encodes the solution of a problem as a ground state of a final hamiltonian . in general form , the hamiltonian that implements a unitary @xmath26 in agt is @xmath131 we call this _ gate hamiltonian _ in our paper . here
, the hamiltonian is defined on bipartite hilbert space @xmath132 and consists of two - body interaction , products of pauli operators rotated by @xmath26 .
this kind of hamiltonian is sometimes referred to as `` twisted hamiltonian '' in @xcite .
this gate hamiltonian is closely related to the heisenberg spin interaction by @xmath133 { \notag \\}&\approx \omega y_j ( { \bm s}_i \cdot { \bm s}_j ) y^\dagger_j , \end{aligned}\ ] ] where @xmath134 represents the direction of the spin and we neglect the identity term because it only shifts the energy spectrum and does not contribute to the adiabatic evolution .
the @xmath135 conjugation term rotates the reference frame @xmath136 around the @xmath137-axis .
the term @xmath138 with positive coupling constant @xmath139 corresponds to a anti - ferromagnetic heisenberg spin interaction .
eq . can be rewritten in the form @xmath140 which may be useful when one considers experimental implementation . to perform agt ,
the following conditions have to be satisfied , 1
. the gate hamiltonian can be implemented .
2 . the ground state of the initial hamiltonian can be prepared .
3 . control the strength of hamiltonian is possible .
i.e switching off the initial hamiltonian and switching on the final hamiltonian can be done gradually .
we assume that these conditions are satisfied in the rest of paper .
we also assume that @xmath141 is a constant resource , and can be used freely .
we introduce the intrinsic computational power of gate hamiltonian .
actually , we can implement various unitary operations @xmath142 and @xmath143 by using the same gate hamiltonian @xmath144 . as we saw in the previous section
, we can teleport the unitary gate @xmath26 to obtain output @xmath145 using hamiltonian eq . .
now , consider the situation when instead of the final hamiltonian @xmath146 , we use @xmath147 i.e. we use the hamiltonian @xmath148 { \notag \\}&= -4\omega u_2^{t } \left [ ( 1-s){\mathinner{|{i}\rangle\!\rangle}}{\mathinner{\langle\!\langle{i}|}}_{23 } + s { \mathinner{|{i}\rangle\!\rangle}}{\mathinner{\langle\!\langle{i}|}}_{13 } \right ] u_2^{t \dagger } .\label{agttrans } { \notag \\}\ ] ] in the last line , we used a property eq . .
this hamiltonian can be written in the unitary conjugation form of adiabatic teleportation from the first qubit to the second qubit ( the third qubit acts as a mediator ) with @xmath149 so we obtain @xmath150 as a output state after the adiabatic evolution .
therefore , the gate @xmath151 can be teleported by just changing the interaction order in the final hamiltonian from @xmath152 to @xmath153 , see fig .
[ agtinitial ] . and @xmath154 by using the same gate hamiltonian in the initial hamiltonian .
] now , let us consider using gate hamiltonian @xmath144 in the final hamiltonian .
we can consider the following two agt schemes .
first possibility is to take the following hamiltonian , @xmath155 where we have used property eq . during the calculation .
therefore , we start with the initial state @xmath66 and get the final state as @xmath156 .
this shows that the result of the adiabatic evolution with hamiltonian eq .
is to implement @xmath157 .
note that we do not need to prepare the ground state of gate hamiltonian @xmath158 in this case .
the remaining possibility is to use the final hamiltonian @xmath159 , i.e. @xmath160 so we will get @xmath161 after agt . and
@xmath162 by using the gate hamiltonian @xmath144 in the final hamiltonian . ]
this shows that @xmath142 , and @xmath143 can be implemented using the gate hamiltonian @xmath163 .
this is summarized in fig .
[ tab - uni ] . [ cols="^,^,^,^,^,^",options="header " , ]
to implement an arbitrary unitary operation on the system , we need to perform consecutive unitaries . in this section , we will consider a case where we perform @xmath164-ordered single - qubit unitary gate teleportations in a single step .
this can be done by small alteration of original agt scheme .
consider @xmath164 gate hamiltonians , @xmath165 @xmath166 .
our aim is to implement @xmath167 operation to the input state @xmath80 in a single adiabatic evolution .
we use @xmath168 qubit system @xmath169 .
initial and final hamiltonian of this parallelized adiabatic gate teleportation ( pagt ) are @xmath170 the main idea comes from just parallelizing agt scheme to achieve consecutive _ time travel _ in a single step , see fig .
[ agtparallel ] .
again , we consider dragging the initial hamiltonian @xmath31 to the final hamiltonian @xmath32 adiabatically , @xmath171 times iteration of unitary gate teleportation in a single adiabatic shift . here
, the initial state @xmath172 is the ground state of the initial hamiltonian eq . ,
and the final state @xmath173 is the ground state of the final hamiltonian eq . ] to see that this adiabatic scheme actually implements a @xmath167 gate teleportation , we start by considering an easier case where all gates are the identity .
i.e. we use @xmath174 we need to check that the input state @xmath112 on @xmath175 teleports to @xmath176 after the evolution .
this fact can be easily proven using the same technique as in the original agt scheme @xcite which uses stabilizer formalism , also explained in section [ subsec : agt ] .
so we consider the stabilizers in the initial hamiltonian eq . as @xmath177 and the final hamiltonian eq . as @xmath178 .
note that there are @xmath168 qubits but the number of generators is @xmath179 , so we have @xmath180 degrees of freedom .
we define logical operators , @xmath181 .
note that @xmath182 or @xmath183 commutes with all stabilizer operators @xmath184 and @xmath185 , i.e. @xmath186=0 $ ] .
so the subspace spanned by these logical operators is preserved throughout the evolution . in the initial stage , \{@xmath187 }
is stabilized , so @xmath188\cdots[x_{2l}x_{2l+1}]=x_1 $ ] and similarly @xmath189 . in the final phase ,
we see that @xmath190 and @xmath191 .
so , the teleportation from the first qubit to the @xmath168-th qubit is faithfully achieved in the ideal case .
now , we consider the general case eq . .
since the identity gate hamiltonian @xmath192 has the property @xmath193 due to eq . , we can write the total hamiltonian in the unitary conjugation form of eq . , @xmath194 where @xmath195 \cdot{\notag \\}\cdots \cdot [ { u^{(1)}_{2l+1}}{u^{(1 ) * } _ { 2l}}\cdots{u^{(1)}_5}{u^{(1 ) * } _ 4}{u^{(1)}_3 } ] ) .\end{aligned}\ ] ] combining these results , we conclude that this parallelized agt hamiltonian @xmath196 teleports the initial state @xmath197 to the final state @xmath198 .
we can perform @xmath164 consecutive unitary operations by iterating agt scheme @xmath164 times , or performing these @xmath164 operations in a single step using pagt .
it is worth pointing out the main differences between agt iteration and pagt .
pagt uses @xmath168 qubit system , and the energy gap @xmath199 depends on @xmath164 .
let us define @xmath200 as the energy gap of an @xmath164-gate pagt scheme , and @xmath201 } \delta e_l ( s)$ ] to be the minimum energy gap .
so , @xmath202 represents the energy gap of usual agt scheme eq . .
we want to consider the scaling behavior of the energy gap @xmath203 .
energy spectrum does not change under unitary conjugation , so it suffices to analyze the case @xmath204 = : h^{s - chain}(s ) .\label{antiferro}\end{aligned}\ ] ] here , the constant identity term is for neglecting @xmath205 term in the hamiltonian , since it does not affect the energy gap .
therefore , @xmath203 behaves like the energy gap of an alternating bond 1-d anti - ferromagnetic spin chain model .
especially when @xmath84 this is an open boundary 1-d spin chain model with size @xmath168 . before calculating the energy gap of eq
, we note that there exists invariant subspaces during the evolution ( see also @xcite ) . the total spin defined by @xmath206 commutes with hamiltonian eq . , namely @xmath207=0 $ ] , we can consider subspace @xmath208 as @xmath209 all @xmath168 subspaces are invariant under the evolution , i.e. @xmath210 . the two degenerate ground states are in the subspace @xmath211 and @xmath212 .
more precisely , @xmath213 and @xmath214 . considering @xmath215 , it only transitions to another state inside @xmath211 during the evolution caused by @xmath216 .
so we only need to consider the energy spectrum in @xmath211 .
same logic is applicable to the state @xmath217 in @xmath212 .
there is symmetry in total spin , the hamiltonian eq .
remains unchange by transforming @xmath218 $ ] .
so the behavior of @xmath219 is the same with @xmath220 , thus it suffices to consider only the energy gap of ground state and first excited state in the subspace @xmath211 . from hereon
@xmath203 denotes the energy gap in this subspace .
we studied this energy gap numerically , by making use of titpack ver.2 which is a library for diagonalizing quantum spin hamiltonians developed by nishimori @xcite .
we have calculated @xmath203 by setting @xmath221 , with @xmath222 for each @xmath164 between @xmath223 to @xmath224 .
the result is displayed in fig .
[ fig : energygap ] . from these results , we see that the minimum energy gap @xmath225 is obtained at @xmath84 .
we plot @xmath225 for each @xmath164 in fig .
[ fig : mine ] , and it can be seen that it scales as @xmath226 . now , we will consider the total time @xmath227 necessary for pagt to be achieved .
when @xmath53 is varied locally on time according to the energy gap @xmath203 , the sufficient time calculated using eq .
is , @xmath228 @xmath229 is also calculated numerically , in fig .
[ fig : te ] .
we have mentioned in in section [ adiabatic theorem ] that @xmath229 scales with @xmath230 in most cases ( e.g. @xcite ) .
it seems that pagt scheme also scales with @xmath231 .
together with the fact that ( see appendix for proof ) @xmath232 we finally get @xmath233 . therefore , we conclude that we need total time @xmath234 to implement consecutive @xmath164-gate operation , while this time scales linearly by iterating the original agt scheme or using quantum circuit model .
we note that our criteria eq . , eq .
are not shown to be a necessary condition for adiabatic theorem , and it may be possible to reduce this total time .
this shows that parallelizing agt does not offer an advantage in terms of the total time for implementing @xmath164 consecutive unitary gates .
however , main advantage of pagt lies in its ability to manipulate causal order by mapping time aligned causal order of gate operations to spatial order of interactions in the final hamiltonian .
the important point of pagt is that the order of the gate teleportations are determined in a single dragging of the hamiltonian .
this order is determined by the form of the final hamiltonian only , and it is not irrelevant to the time parameter @xmath53 .
let us consider the case @xmath235 . in the above pagt scheme ,
@xmath236 will be applied to the target state . however , if the final hamiltonian is changed to @xmath237 rather than using eq .
, one can easily check that the output state after this pagt becomes @xmath238 .
the pagt scheme enables us to encode information about unitary gates and their order _
separately_. i.e. unitary gates are encoded in the initial hamiltonian , and their order is determined by the final hamiltonian .
this model enables us to map causal order of operation into spatial interaction between qubits .
this interesting property enables us to control order with control - qubit which we will explain in the next section .
so far we have introduced the parallelized version of agt , which increases the total time compared to iterative agt .
however , we now show how different pagt evolutions with the same initial hamiltonian can be actually superposed using a control - qubit .
this controlled - pagt allows us to manipulate the causal order of gate operations by a control - qubit and implement a quantum switch ( qs ) , first introduced in @xcite .
assume a two - qubit system consisting of a control - qubit and a target qubit , described by @xmath239 .
we also assume we can use two single - qubit unitary gates , f and g @xmath240 .
qs is a super operator ( one may regard it as black box itself ) , @xmath241 .
it takes @xmath2 and @xmath3 , and outputs the unitary operation @xmath242 .
more specifically , its action to a state is described by @xmath243 { \notag \\}&={| { 0 } \rangle}_c gf{| { \phi_0 } \rangle}_t + { | { 1 } \rangle}_c fg{| { \phi_1 } \rangle}_t .\label{qs}\end{aligned}\ ] ] @xmath244 are the ket vectors for the target state and are not normalized .
when control - qubit is in the @xmath245 state , unitary operation @xmath2 acts first on the target qubit followed by @xmath3 .
and if the control - qubit is @xmath246 state , the order of operation changes so @xmath3 acts first followed by @xmath2 .
this quantum control preserves coherence . rather than calling @xmath2 and @xmath3 as `` unitary gates '' ,
they are referred to as `` oracles '' due to existence of no - cloning theorem for boxes @xcite .
so we will consider now @xmath2 and @xmath3 as oracles rather than elementary gates .
it was shown in @xcite that it is impossible to implement qs using only one call of each unitary oracle @xmath247 in the quantum circuit model with definite causal structure . the same output of eq .
actually can be produced ( see fig . [
fig : qsbyqc ] ) , but it requires the use of either unitary at least twice , which is forbidden in our setting . even if quantum circuit model can not implement quantum switch , the assumption of _ no definite causal structure _ itself is not prohibited by the axioms of quantum mechanics . it was also suggested that if one can somehow implement _ superposition of a quantum wire _ in quantum circuit , we can implement this operation , although the concrete physical system to achieve these kind of superposition has not been suggested yet .
implementation of qs operation in a quantum circuit is possible if one uses measurement and allows probabilistic success . in the following ,
we show how the transformation can be simulated by pagt . here
, we use the term `` simulate '' , because we are using a different resource , namely the gate hamiltonian @xmath248 to implement unitary operation @xmath247 .
we consider a system consisting of six qubits described by @xmath249 .
the control - qubit is encoded in @xmath250 and the target qubit is encoded in @xmath251 .
we also assume the gate hamiltonians are @xmath252 and @xmath253 , corresponding to unitary gates @xmath2 and @xmath3 .
we prepare the input state @xmath254 on the system @xmath255 .
initial hamiltonian is written as @xmath256 note that @xmath250 and @xmath9 is free with eq . , it has four - fold degenerate ground state .
the initial state is written in the form , @xmath257 the main idea comes from the fact that we can manipulate the order of operations by the final hamiltonian in pagt .
we design the final hamiltonian to control the order of unitary gates by the control - qubit as follows , @xmath258 we construct the different interactions described by @xmath259 and @xmath260 through adiabatic process depending on the control - qubit . our intuition comes from adiabatically simulating fig .
[ figagtqs ] .
the total hamiltonian is @xmath261 this hamiltonian implements the operation implemented by qs , @xmath262 , as we prove below
. we can restrict our attention to the most simple case when @xmath263 ( we write corresponding hamiltonian @xmath264 ) , because implementing unitary @xmath247 can be written in the unitary conjugation form @xmath265 therefore , considering this simplest case is enough to understand the general evolution .
we prove the achievability of adiabatic teleportation from @xmath254 to @xmath266 .
when using stabilizer formalism to show this , logical stabilizer of the target qubit can be considered as @xmath267 which commute with @xmath268 . in the same way , we can define logical @xmath60 in control qubit @xmath269 since it commutes with @xmath270 .
however when we try to define logical x as @xmath271 , it does not commute with @xmath270 .
so explanation by this approach now fails and we need to consider the dynamics explicitly .
. can be written , @xmath272 { \notag \\}&+{{| { 1 } \rangle } { \langle { 1 } |}}_c \otimes \left [ ( 1-s)h^{qsat}_{ini ~/c } + s h_{fin}^{(1 ) } \right].\end{aligned}\ ] ] note that @xmath33 is a function of @xmath34 where @xmath273 is the hamiltonian where the control system @xmath250 is subtracted from @xmath274 .
we will define @xmath275 .
let us write the initial state as @xmath276 .
time evolution operator is written by @xmath277 where @xmath278 is a time - ordering operator , thus @xmath279 the state time evolution @xmath280 is the same when one consider previous pagt scheme with @xmath235 .
after the process , we have @xmath281 however , we must now consider the relative phase between these states .
we want to know the condition that ensures the relative phase remains 0 . to analyze it
specifically , we will consider the following time - dependent hamiltonian @xmath282.\end{aligned}\ ] ] here we assume the situation where the hamiltonian strength @xmath33 is not controlled simultaneously each other ( this consideration comes from the experimental setting where we may not be able to synchronize the strength @xmath33 of each term ) . for simplicity we also set @xmath283 in the following , @xmath284 relabelling the hilbert space of @xmath285 to @xmath286 and @xmath287 , we get @xmath288 comparing with eq . , we know @xmath289 are the sufficient condition that the relative phase between @xmath290 and @xmath291 remains 0 .
i.e. to achieve controlled adiabatic teleportation , we restrict each of the strength of 1 .
gate hamiltonian @xmath252 and @xmath253 2 .
first gate between control @xmath245 and @xmath246 3 .
second gate between control @xmath245 and @xmath246 are exactly the same for arbitrary time @xmath34 through the evolution . especially , in the case that the strength of each hamiltonian is controlled in the same manner @xmath53 as in eq . , it suffices the condition .
this shows that controlled teleportation works with condition eq . .
here , we note that one of the good property in agt losts with this scheme . in agt , precise time control
is not required for implementing unitary gate because we can implement @xmath26 to an unknown sate @xmath80 only if the time parameter @xmath53 is slow enough .
however , now we need to synchronize the strength of @xmath53 to satisfy eq . during the evolution . from this analysis
, we can see that the state after the evolution caused by general form eq .
is @xmath292 followed by c - swap ( controlled - swap ) operations , we can confirm that the quantum switching operation is achieved @xmath293 note again that we are not using the oracle @xmath2 and @xmath3 , but gate hamiltonian @xmath252 and @xmath253 . if eq .
is satisfied , it is easy to understand that the energy eigenvalues do not change from pagt scheme .
we denote the eigenvalues and the eigenvectors of @xmath294 as @xmath295 and @xmath296@xmath297 and @xmath298 , @xmath299 this allows us to find the eigenvalues and the eigenvectors of controlled version eq . , @xmath300 since the eigenvalues remain unchanged from pagt scheme eq . , agt of the quantum switch is possible .
we stress that we control the final hamiltonian which decides only the order of operations , and it does not depend on gate hamiltonian .
in quantum circuit model it is straight forward to count the number of oracle , however in adiabatic model , `` number of calls '' does not have well - defined meaning .
therefore , fair comparison between 2 models of resources is not straight forward . here
, we will introduce the term `` dependent resource '' and `` independent resource '' instead of the notion `` once '' and `` twice ( or more times ) '' respectively for our purposes .
for example , quantum circuit in fig . [ fig : qsbyqc ] produces the same result of qs , however it uses @xmath2 twice .
we will denote them as @xmath301 and @xmath302 .
then we say that these @xmath301 and @xmath302 are independent resource , because these operation act on different spacetime , i.e. they act either at different time or different qubits .
even if the resource used to produce the unitary @xmath2 is the same , we call it `` independent resource '' if it is used at different times
. we call them `` independent '' because we can only change @xmath301 operation without changing @xmath302 , or let us assume that there is noise on @xmath301 this affects only @xmath301 and @xmath302 does not affected . and @xmath302 . here
, a combination of @xmath303 represents swap operation .
first two swap operations with black circle is performed when control - qubit is @xmath246 , and last two swap operations with white circle is performed when control - qubit is @xmath245 . ]
contrary to this model , our controlled - pagt scheme uses the dependent resources @xmath304 to implement @xmath305 operations . here ,
both @xmath301 and @xmath302 are derived from the same resource @xmath252 , and both @xmath306 and @xmath307 are derived from @xmath253 .
thus , we regard each @xmath252 and @xmath253 is dependent resource to implement @xmath2 and @xmath3 unitary operation in qs operation .
to summarize , we rephrase the result in @xcite and conclude as follows . _ in definite causal structure , quantum switch can not be implemented by quantum circuit model without using independent resource deterministically .
however , controlled - pagt can simulate this task only using dependent resource .
_ there are some papers which studies about the computational power of quantum switch @xcite .
the definition of qs can be generalized so that @xmath2 and @xmath3 are quantum channels , cptp maps .
it was proven that the power of generalized qs has a power beyond quantum circuit .
more precisely , one can achieve a channel discrimination task which can not be achieved with usual standard quantum circuit @xcite .
the resource comparison of the order controlled operation implemented by qs and standard quantum circuit was performed in @xcite .
however , we note that there is no computational speed up if one can use the oracle many times . because we can implement quantum switch in quantum circuit in polynomial time .
we will consider a question of implementing controlled - unitary using our controlled - pagt technique .
our task is to implement controlled-@xmath26 gate from given unitary @xmath26 .
this question is important because some well - known quantum algorithms require us to call a controlled - oracle .
phase estimation algorithm is one such example , which has important applications that include shor s factoring algorithm @xcite .
however , it was shown by soeda @xcite that it is not possible to implement controlled - unitary if one can only use an uncontrolled oracle . in this section
, we consider the implementation of a controlled - unitary and show that it is possible in certain case , where the oracle is written in orthogonal matrix form .
we consider the problem of implementing a controlled-@xmath26 using an gate hamiltonian @xmath308 .
let us consider the following hamiltonian which is defined on the hilbert space @xmath309 , @xmath310 we assume the input state is @xmath311 . see fig .
[ agtcu ] .
simple analysis shows that it does not implement the controlled-@xmath26 operation because after the scheme , we get @xmath312 so even after c - swap , ancilla state on @xmath313 has incoherent part .
therefore , when we only look at output state , space on @xmath314 in eq . , it becomes a mixed state .
thus this scheme fails to implement a controlled - unitary .
previous consideration says that we need to _ erase _ the information of @xmath158 in the @xmath315 part .
therefore , we consider the following scheme @xmath316 we add the term @xmath317 to shift the initial state @xmath318 to the final state @xmath319 .
however , this scheme generates another problem .
when we focus on the evolution from @xmath318 to @xmath319 by adiabatic process , it depends on the form of @xmath158 .
for example , in the case @xmath320 , there is an energy crossing between ground state and first excited state during the evolution , in which case the deletion becomes impossible to perform . also , even @xmath158 can be deleted , the final state become @xmath321 } { | { 0 } \rangle}_c { \mathinner{|{i}\rangle\!\rangle}}_{12 } { \mathinner{|{i}\rangle\!\rangle}}_{45 } { | { \phi_0 } \rangle}_3 + { | { 1 } \rangle}_c { \mathinner{|{i}\rangle\!\rangle}}_{12 } { \mathinner{|{i}\rangle\!\rangle}}_{45 } u{| { \phi_1 } \rangle}_3 \label{cuoutrevised}.\end{aligned}\ ] ] here , extra relative phase factor @xmath322}$ ] arises in the @xmath315 part . @xmath323 $ ] is a functional which depends on the form of the function @xmath53 . in general
, we do not extract the value of @xmath324 . in this case
, one has to average over the states , which gives the following density operator @xmath325
so , the coherence between @xmath315 and @xmath326 is broken .
the same density operator is obtained when one measure control - qubit on the @xmath327 basis , followed by operating unitary @xmath26 when one get the measurement result @xmath246 .
this scheme does not implement the controlled - unitary either .
the above protocol does not work well . however , let us remind the power of gate hamiltonian we explained in section [ sec : gateham ] .
we can implement @xmath26 and @xmath154 by using same gate hamiltonian @xmath144 .
now , let us assume that @xmath328 where @xmath329 is an orthogonal matrix , i.e. @xmath330 .
we consider following hamiltonian which is defined on the hilbert space @xmath309 , @xmath331 is implemented by using gate hamiltonian @xmath332 ] when control - qubit is @xmath245 , the gate teleportation acts @xmath333 .
therefore , the output state is @xmath334 indeed , controlled - orthogonal unitary operation is implemented .
therefore , we can implement controlled-@xmath335 operations from gate hamiltonian @xmath336 by controlled - pagt scheme .
indeed , we can also consider stranger operations , such as the process showed in fig .
[ agtutd ] implements a controlled-@xmath337 operation .
operation by using gate hamiltonian @xmath144 in controlled - pagt scheme . ] in this chapter , we show an application of controlled - pagt to implement controlled version of an unknown unitary gate .
however , these applications are not using coherent control of the causal order of gate operations .
in fact , the possibility of the tasks is based on the power of gate hamiltonian @xmath144 , which can implement @xmath338 and @xmath143 operations .
therefore , this shows the difference of the power between gate hamiltonian @xmath144 in agt model , and unitary gate @xmath26 in standard quantum circuit model .
we have investigated a method to manipulate the causal order of gate operations by adiabatically implementing quantum computation .
our method is based on the adiabatic gate teleportation ( agt ) scheme proposed by bacon and flammia @xcite .
we introduce _ gate hamiltonian _ as a resource to implement a unitary operation .
we have shown that this resource can implement not only the unitary but also the transpose , complex conjugate , and adjoint of the unitary operation . using the gate hamiltonians ,
we construct a parallelized agt ( pagt ) scheme , where consecutive gate operations are implemented in a single adiabatic shift . an important feature of the pagt scheme is that information about the unitary gate and information about the causal order of gates are separately encoded in the initial hamiltonian and the final hamiltonian respectively .
we can choose the total unitary gate by changing the gate hamiltonian in the initial hamiltonian , and we can choose the order of these unitary gates by changing the spatial interaction order of the final hamiltonian .
this property allows us to construct the controlled - pagt scheme that uses a control - qubit to manipulate the causal order of the unitary gates . however , pagt has no advantage in terms of the computational time required to implement @xmath164 consecutive gate operations compared to agt scheme since the energy gap narrows by parallelization .
quantitatively , total time scales with @xmath234 to implement @xmath164 consecutive unitary gate operations in pagt , while it scales @xmath339 by just iterating the agt scheme for @xmath164 times . 0
the crucial difference obtained by parallelizing agt is that pagt maps the causal order of unitary operations to spatial order , and we can even control this order with control - qubit to construct controlled - pagt . this advantage is crucial in the controlled - pagt that uses a control - qubit to manipulate the order of the unitary gates . in controlled - pagt ,
we can perform coherently order controlled unitary operation , which is equivalent to the operation implemented by quantum switch .
this result indicates the difference between pagt model and standard quantum circuit model in terms of `` causal order manipulation '' .
but we note that even one considers order controlled operation by controlled - pagt , same unitary operation can be simulated by quantum circuit within time @xmath339 once we allow to use unitary gate many times .
it is a open problem that whether manipulating causal order contribute some computational speed up .
there are some researches about `` acausal operations '' recently , it was shown that mbqc with g - flow can be translated into quantum circuit with acausal loops @xcite , and the conditions for this quantum circuit with acausal loops can be decomposed into usual quantum circuit is studied in @xcite . also , the correspondence between quantum computational power and causal order in locc setting is studied in @xcite .
we would like to acknowledge elham kashefi for introducing adiabatic gate teleportation .
we also would like to thank shunsuke furukawa for helpful discussions .
this work is supported by the project for developing innovation systems of mext , japan , the global coe program of mext japan , and jsps kakenhi ( grant no .
23540463 , and no .
23240001 ) .
we also gratefully acknowledge to the elc project ( grant - in - aid for scientific research on innovative areas mext kakenhi ( grant no .
24106009 ) ) for encouraging the research presented in this paper .
here , we prove eq . by considering lower bound and upper bound of this equation .
lower bound is obtained by using the property , @xmath340 is arbitrary chosen vector . in our case , choosing @xmath341 as @xmath342 shows that latexmath:[\[\begin{aligned } d. aharonov _ et al .
_ , eprint arxiv : quant - ph/0405098 ( 2004 ) . b. w. reichardt , in _ proceedings of the thirty - sixth annual acm symposium on theory of computing _ , stoc 04 , pp .
502510 , new york , ny , usa , 2004 , acm . | we study adiabatic gate teleportation ( agt ) , a model of quantum computation recently proposed by d. bacon and s. t. flammia , phys .
rev . lett . * 103 * , 120504 ( 2009 ) , to investigate a new property of quantum computation , namely causal order manipulation .
we develop _ parallelized adiabatic gate teleportation _ ( pagt ) where a sequence of gate operations is performed in a single step of the adiabatic process by introducing a _ gate hamiltonian _ implementing a single - qubit unitary gate . by parallelizing the agt scheme , the necessary time for the adiabatic evolution implementing a sequence of gates increases quadratically
, however it allows us to map causal order of gate operations to spatial order of interactions in the final hamiltonian . using this property ,
we also present controlled - pagt scheme to manipulate the causal order of gate operations by a control - qubit .
the scheme allows us to coherently perform two differently ordered unitary operations @xmath0 and @xmath1 depending on the state of a control - qubit by simultaneously applying the gate hamiltonians of @xmath2 and @xmath3 . |
the role of the gluon field in determining the interactions between nucleons and quark - antiquark ( @xmath12 ) systems , which do not share the same @xmath13 and @xmath14 quarks with the nucleon , is one of the interesting subjects in understanding quantum chromodynamics(qcd ) . an important step toward this direction
was taken by peskin@xcite who applied the methodology of the operator product expansion to evaluate the strength of the color field emitted by heavy @xmath12 systems .
his results suggested@xcite that the van der waals force induced by the color field of @xmath0 on nucleons can generate an attractive short - range @xmath0-@xmath15 interaction . by using the effective field theory method ,
luke , manohar , and savage@xcite used the results from peskin to predict the @xmath0-nucleon forward scattering amplitude which was used to get an estimation that @xmath0 can have a few mev / nucleon attraction in nuclear matter .
brodsky and miller@xcite further investigated the @xmath0-n forward scattering amplitude of ref.@xcite to derive a @xmath0-n potential ( @xmath5 ) which gives an @xmath0-n scattering length of -@xmath16 fm .
the result of peskin was also used by kaidalov and volkovitsky@xcite , who differed from ref.@xcite in evaluating the gluon content in the nucleon , to give a much smaller scattering length of -@xmath17 fm . in a lattice qcd calculation ,
kawanai and saski@xcite obtained an attractive @xmath0-n potential @xmath18 - @xmath19 with @xmath20 and @xmath21 gev , which gives a scattering length - @xmath22 fm . in ref.@xcite ,
brodsky , schmidt and de teramond proposed an approach to calculate the potential between a @xmath23 meson and a nucleus by using the pomeon - exchange model of dannachie and lanshoff@xcite .
the @xmath0-n potential obtained in this approach is @xmath18 - @xmath24 with @xmath25 and @xmath21 gev which gives a rather large scattering length - @xmath26 fm .
our first objective in this paper is to explore whether these @xmath0-n potentials , with rather different attractive strengths , can form @xmath0-nucleus bound states . following the well developed method in nuclear reaction theory@xcite ,
this is done by searching for bound states by solving the schrodinger equation with a folding potential constructed by integrating the @xmath0-n potential over the nuclear density .
we will also consider the approach of ref.@xcite in predicting @xmath0-nucleus bound states by the coherent sum of pomeon - exchange between quarks in @xmath0 and all quarks in the nucleus .
for each of the predicted bound @xmath2_{j/\psi}$ ] systems , we then estimate the photo - production cross section of the @xmath27_{j/\psi } + n$ ] reaction to facilitate future experimental investigations@xcite .
the second part of this work is motivated by the investigations by brodsky and de teramond@xcite who found that the spin correlation of @xmath28 elastic scattering near the @xmath0 production threshold can be explained if one postulates the excitation of a hidden charm ( @xmath29 ) state @xmath30 .
based on the similar consideration on the role of multi - quark configurations , brodsky , chudakov , hoyer , and laget@xcite suggested in a study of @xmath31 reaction that @xmath0 can interact strongly with the six - quark @xmath7 $ ] component of the deuteron wavefunction because the octet 3-quark @xmath32_8 $ ] in the @xmath7 $ ] can directly interact with each quark in @xmath0 .
these two works suggest the possibility that if @xmath0 overlap with a @xmath7 $ ] cluster in nuclei , a bound @xmath7_{j/\psi}$ ] system could be formed .
it is of course very difficult , if not impossible , to estimate @xmath7$]-@xmath33 interaction .
instead , we will simply assume the existence of such states and use the previous works@xcite on quark clusters in nuclei to explore how the cross sections of @xmath9_{j/\psi } + n$ ] depend on the parameters characterizing the @xmath7$]-@xmath33 interaction within a potential model .
our first task is to construct a model of @xmath34 reaction . at high energies
, it is well recognized that this reaction can be described by the pomeron - exchange model with an interpretation@xcite that pomeron - exchange is due to the exchange of gluons within qcd .
this is illustrated in part ( a ) of fig.[fig : pom - mec ] . at low energies ,
one expects that mechanisms other than pomeron - exchange could also contribute as can be seen in the exclusive @xmath35 photo - production reaction on the nucleon@xcite .
however , very little investigation has been done for @xmath0 photo - production in the near threshold region . as a first step , we will only consider the meson - exchange mechanism which can be calculated from using the partial decay width of @xmath36 listed by particle data group@xcite ( pdg ) . with the vector meson dominance ( vdm ) assumption , this observed decay process indicates that @xmath0 photo - production can also be due to the exchanges of a @xmath37 meson with the nucleon , as illustrated in part ( b ) of fig.[fig : pom - mec ] .
we next consider the photo - nuclear reaction mechanism that a @xmath0 is produced from a nucleon in a nucleus with mass number @xmath38 and then forms a bound state with the spectator @xmath39 system which can be a @xmath40 nuclear system or a quark cluster @xmath41 $ ] in the target nucleus @xmath38 . with this commonly used impulse approximation ,
the reaction cross sections can be calculated from the @xmath34 amplitude , which will be generated from the pomeron - exchange and pion - exchange mechanisms described above , and the initial nucleon and final @xmath0 wavefunctions .
the nucleon wavefunctions can be taken from the available nuclear models .
the @xmath0 wavefunctions will be generated from various @xmath0-@xmath39 potentials mentioned above . for simplicity
, we only present the predictions of the cross sections of @xmath42 \rightarrow n + [ ^3he]_{j/\psi } $ ] and @xmath43 \rightarrow n + [ q^6]_{j/\psi } $ ] reactions . in section
ii , we present formula for calculating the @xmath34 amplitudes from the pomeron - exchange and pion - exchange mechanisms .
the impulse approximation formula for calculating the cross sections of @xmath44 \rightarrow [ b]_{j/\psi } + n$ ] are given in section iii .
our results are presented in section iv . in section
v , we give a summary and discuss the necessary future work .
in the center of mass system , the differential cross section of @xmath45 with invariant mass @xmath46 can be written @xmath47[\frac{ke_{j/\psi}(k)e_n(k)}{w } ] \frac{1}{4}\sum_{\lambda_\gamma,\lambda_{j/\psi}}\sum_{m_s , m'_s } where @xmath48 and @xmath49 are the helicities of the @xmath0 and photon , respectively , @xmath50 is the z - component of the nucleon spin , and @xmath51^{1/2}$ ] is the energy of a particle with mas @xmath52 .
the reaction amplitude is written as @xmath53\ , , \label{eq : tot - amp}\end{aligned}\ ] ] where @xmath54 is the nucleon spinor ( with the normaliztion @xmath55 ) , @xmath56 and @xmath57 are the polarization vectors of @xmath0 and photon , respectively . here
we also have introduced the four - momenta for the initial and final nucleons : @xmath58 in the following subsections , we give formula for calculating the invariant amplitude @xmath59 due to the pomeron - exchange and meson - exchange mechanisms , as illustrated in fig.[fig : pom - mec ] . within the pomeron - exchange model of donnachie and landshoff @xcite ,
the vector meson photo - production at high energies is due to the mechanism that the incoming photon couples with a @xmath60 pair which interacts with the nucleon by the pomeron exchange before forming the outgoing vector meson .
the quark - pomeron vertex is obtained by the pomeron - photon analogy@xcite , which treats the pomeron as a @xmath61 isoscalar photon , as suggested by a study of non perturbative two - gluon exchanges @xcite . following the formula given explicitly in ref.@xcite , we then have @xmath62 with @xmath63 where @xmath64 , @xmath65 , @xmath66 , @xmath67 is the pomeron - quark coupling constant , @xmath68 is the vector meson mass , and @xmath69 is the isoscalar electromagnetic form factor of the nucleon , @xmath70 here @xmath71 is in unit of gev@xmath72 , and @xmath73 is the proton mass . the regge propagator for the pomeron in eq .
( [ eq : mp ] ) is @xmath74 \right\ } \ , , \label{eq : regge - g}\ ] ] where @xmath75 .
it is common@xcite to use @xmath76 and @xmath77 gev@xmath78 . in eq .
( [ eq : pom ] ) , @xmath79 is the vector meson decay constant : @xmath80 , @xmath81 , @xmath82 , and @xmath83 .
the other parameters in eq.([eq : pom ] ) have been determined by fitting@xcite the total cross section data of the photo - production of @xmath84 and @xmath85 : @xmath86 gev@xmath87 and @xmath88 gev@xmath72 . with the parameters specified above , our task is to examine the extent to which the total cross section of photo - production of @xmath0 can be fitted by only adjusting the pomeron - charmed quark coupling constant @xmath89 .
this will be discussed in section iv .
we observe from particle data@xcite that the width of the @xmath90 is significant , @xmath91 with the vector meson dominance ( vdm ) assumption , this experimental information allows us to calculate the one - pion - exchange amplitude of @xmath34 , as illustrated in fig.[fig : pom - mec](b ) , by using the following lagrangian @xmath92 with @xmath93 where @xmath94 , @xmath95 , @xmath96 , @xmath97 , and @xmath98 are the field operators for @xmath99 , @xmath0 , @xmath37 , photon ( @xmath100 ) , and nucleon ( @xmath15 ) , respectively .
the mass for particle @xmath101 is denoted as @xmath52 .
the well determined coupling constants are @xmath102 , @xmath103 , and @xmath104
. to determine @xmath105 , we use @xmath106 given in eq.([eq : l - jrp ] ) to calculate the decay width @xmath107 where @xmath108 is defined by @xmath109 , and @xmath110 ^ 2\ , .
\label{eq : width-2}\end{aligned}\ ] ] here we have included a dipole cutoff function with a range parameter @xmath111 .
the four - momenta are defined in the rest frame of @xmath0 : @xmath112 by using eqs.([eq : width-1])-([eq : width-2 ] ) and the experimental value given in eq.([eq : jpsi - width ] ) , we find @xmath113 for a cutoff @xmath114 mev . with the lagrangian eq.([eq : larg ] ) , the one - pion - exchange invariant amplitude for @xmath115 can be written as @xmath116 with @xmath117\ , , \label{eq : opep}\end{aligned}\ ] ] where @xmath64 , and @xmath118 here we have introduced a cutoff form factor @xmath119 to regularize the interaction verteces . for simplicity ,
we use the following form @xmath120 we set @xmath121 , and @xmath122 mev .
with the impulse approximation , we assume that a @xmath0 is produced on a nucleon in the target nucleus @xmath38 and then is attracted by a spectator system @xmath39 to form a bound state @xmath123_{j/\psi}$ ] . for simplicity , @xmath123_{j/\psi}$ ]
is denoted as @xmath124 in the following formula . with the mechanism illustrated in fig.[fig : ghe4 ]
, the cross section of @xmath125 in the center of mass system can be written as @xmath126 [ \frac{pe_n(p)e_d(p)}{w } ] \nonumber \\ & & \times \frac{1}{2}\frac{1}{2j_a+1}\sum_{\lambda , m_{j_a}}\sum_{m_s , m_d } \label{eq : d - prod-1}\end{aligned}\ ] ] where @xmath127 here @xmath128 is the creation operator for a @xmath0 with wavefunction @xmath129 , @xmath130 an annihilation operator for a nucleon with wavefunction @xmath131 , and the @xmath132 amplitude is @xmath133\ , , \nonumber \\ \label{eq : t - mx}\end{aligned}\ ] ] where @xmath134 is the pion - exchange amplitude given in eq.([eq : opep ] ) , and @xmath135 is the pomeron - exchange amplitude in eq.([eq : mp ] ) . for simplicity
, we will only perform calculations for the reactions on @xmath1 and @xmath3 . for estimations of cross sections on these target nuclei , it is sufficient to use the s - wave harmonic oscillator wavefunctions for both the target @xmath38 and @xmath39 in the @xmath123_{j/\psi}$ ] bound state .
we also only consider the case that the @xmath0 in the produced bound @xmath136 is on an s - wave orbital .
for the case @xmath137 nuclear system , we thus write the initial ( @xmath138 ) and final ( @xmath139 ) nuclear states as @xmath140_{l=0 } \,,\label{eq : a - wf}\\ where @xmath141 is the relative angular momentum between @xmath15 or @xmath0 and the @xmath40 nucleus .
explicitly , we have @xmath142 then the momentum variables in eqs.([eq : d - prod-2 ] ) and ( [ eq : t - mx ] ) are @xmath143 where @xmath144 ( @xmath145 ) is the relativistic relative momentum between @xmath0 ( @xmath146 and the @xmath40 nuclear system . for the target @xmath3 , we have @xmath147 and assume that @xmath148 is the @xmath1 ground state with @xmath149 .
we then have the following simplicities : @xmath150 and @xmath151 eq.([eq : d - prod-2 ] ) then becomes @xmath152\ , .
\label{eq : d - prod-3}\end{aligned}\ ] ] we have applied the formula eqs.([eq : d - prod-1 ] ) and ( [ eq : d - prod-3 ] ) to estimate the production cross section on @xmath3 .
we use the usual s - wave harmonic oscillator wavefunction with @xmath153 fm for the target @xmath3 @xmath154\end{aligned}\ ] ] with the normalization @xmath155 . for the @xmath0 wavefunction in @xmath156_{j/\psi}$ ] , we will generate a s - wave @xmath157 from a potential @xmath158 with the normalization @xmath159 .
the wavefunction in eq.([eq : wf - jpsi-0 ] ) and also in ( [ eq : d - prod-3 ] ) can then be calculated from @xmath160 where @xmath161 is the spherical bessel function .
the form of @xmath158 will be discussed in section iv .
the above formula can be easily extended to investigate other possible impulse approximation mechanisms as far as all wavefucnctions in the bound @xmath38 and @xmath123_{j/\psi}$ ] are all in s waves .
this is what we will need in section iv when we consider the production of @xmath7_{j/\psi}$ ] from the @xmath162-n component of @xmath1 .
we first develop a model consisting of pomeron - exchange and pion - exchange mechanisms , as described in section ii . with the parameters specified there , we try to fit the available total cross section data of @xmath164 up to invariant mass @xmath165 gev by only adjusting the charmed quark - pomeron coupling constant @xmath89 . with @xmath166
we only able to fit the data up to 20 gev .
clearly , the result at high energy is not satisfactory as shown in the red dashed curve in the left - hand side of fig.[fg : jpsiphoto ] .
we then find that by changing @xmath167 of the regge trajectory in the pomeron propagator eq.([eq : regge - g ] ) from @xmath168 , as determined in the previous fits@xcite to the total cross sections of @xmath84 and @xmath85 photo - production , to @xmath169 , we are able to get a very good fit to the data by choosing @xmath170 gev@xmath87 . our fit is the solid black curve in the left - hand side of fig.[fg : jpsiphoto ] .
we thus will use the model with @xmath169 and @xmath170 gev@xmath87 ( pm model ) in our investigations . as also seen in the insert in the left - hand side of fig.[fg : jpsiphoto ] , the contribution ( magenta dotted curve ) from the pion - exchange amplitude , as defined by eqs.([eq : mex - amp])-([eq : mex - ff ] ) , is very weak except in the very near threshold region .
reaction as function of the @xmath171 invariant mass @xmath46 . left : the black solid and red dashed are the results from using ( @xmath169 , @xmath172 gev@xmath87 ; called the pm model ) and ( @xmath168 , @xmath173 gev@xmath87 ) within the model which include both the @xmath37-exchange and pomeron - exchange mechanisms .
the magenta dotted curve in the insert is the contribution from the @xmath37 exchange .
right : the black solid , blue dotted and green dot - dashed lines are from the pm model , the @xmath174 model of ref.@xcite , and the @xmath175 model based on eq.([tmatrix ] ) .
the experimental data are from @xcite.,title="fig : " ] reaction as function of the @xmath171 invariant mass @xmath46 .
left : the black solid and red dashed are the results from using ( @xmath169 , @xmath172 gev@xmath87 ; called the pm model ) and ( @xmath168 , @xmath173 gev@xmath87 ) within the model which include both the @xmath37-exchange and pomeron - exchange mechanisms .
the magenta dotted curve in the insert is the contribution from the @xmath37 exchange .
right : the black solid , blue dotted and green dot - dashed lines are from the pm model , the @xmath174 model of ref.@xcite , and the @xmath175 model based on eq.([tmatrix ] ) .
the experimental data are from @xcite.,title="fig : " ] we next consider the model of ref.@xcite based on the two - gluon ( @xmath174 ) and three - gluon ( @xmath176 ) exchange mechanisms . in terms of the normalization
defined by eq.([eq : tot - amp ] ) , the amplitude of this model can be written as @xmath177 \label{tmatrix}\end{aligned}\ ] ] with @xmath178 where @xmath179 fm , @xmath180 gev@xmath78 are taken from ref .
we follow ref.@xcite to determine the parameters @xmath181 and @xmath182 by fitting the data up to only 20 gev . in the two - gluon - exchange model ( @xmath174 ) , we set @xmath183 and obtain @xmath184 mev@xmath78 from the fit . in the @xmath175 model ,
the fit is obtained by choosing @xmath185 mev@xmath78 and @xmath186 mev@xmath78 . the fits for the @xmath174 and
@xmath175 models are the dotted and dot - dashed curves in the right - hand side of fig.[fg : jpsiphoto ] , respectively .
clearly , they have differences with that ( black solid ) of the pm model , as can be seen more clearly in the insert in the right - hand side of fig.[fg : jpsiphoto ] . here
we also see that the data in the region near the @xmath0 production threshold are very limited and uncertain .
we will therefore perform calculations using the pm , @xmath174 , and @xmath175 models to examine the model dependence of our predictions .
clearly , precise data in the near threshold region are needed to make progress . following the previous investigations @xcite , we assume that the interaction between a @xmath0 and a nucleus with mass number @xmath38 can be parameterized as a non - relativistic potential of the following yukawa form @xmath187 there exists two different approaches to determine the parameters @xmath188 and @xmath189 for the nucleon with @xmath190
. we will explain these in the following two subsections .
motivated by the previous studies in quantum electrodynamics , it is assumed in the approach of ref.@xcite that the @xmath0-a forward angle scattering amplitude at very high energy can be related to the matrix element of the potential eq.([eq : yukawa - n ] ) which is understood to be valid only in the region where @xmath0 moves non - relativistically .
they further assume that the @xmath0-a amplitudes can be calculated by using the pomeron - exchange model of dannachie and landshoff@xcite .
in the very high energy approximation , the differential cross section of @xmath0-a elastic scattering can be related to the parameters @xmath191 and @xmath189 of the potential eq.([eq : yukawa - n ] ) by following relation @xmath192 ^ 2 [ 3a\beta_{u / d } f_{a}(t)]^2}{4\pi } \label{eq : dl - jpsi-1 } \\ & = & \frac{4\pi\alpha^2_a}{(-t+\mu^2_a)^2}\ , , \label{eq : dl - jpsi-2}\end{aligned}\ ] ] where @xmath71 is the momentum - transfer squared , @xmath193 ( @xmath89 ) is the pomeron coupling with the @xmath194 ( @xmath195 ) quarks , @xmath196 and @xmath197 are the form factors for @xmath0 and the nucleus with mass number @xmath38 , respectively .
they further assume that in the @xmath198 limit , the slope of @xmath199 is mainly determined by @xmath200 and that @xmath197 can be identified with the nuclear electromagnetic form factor .
one then gets the following relations @xmath201[3a\beta_{u / d}]}{4\pi}\mu^2_a\ , .
\label{eq : alpha}\end{aligned}\ ] ] the radius @xmath202 can be taken from ref.@xcite .
the pomeron - quark coupling constants can be taken from fits to the data of meson - nucleon scattering or photo - production of vector mesons .
once @xmath191 and @xmath189 of the potential eq.([eq : yukawa - n ] ) are determined , we can predict the possible @xmath0-nucleus bound states . in table [ tab : bound-1 ] , we list our results for proton ( @xmath203 , @xmath1 ( a=3 ) and @xmath204 ( a=12 ) for various sets of pomeron - quark coupling constants .
the first rows in the results for each @xmath38 are based on the flavor independent @xmath205 gev@xmath87 of ref.@xcite .
the other two results use the pomeron - quark coupling constants @xmath206 gev@xmath87 determined@xcite in the fits to the data of photo - production of @xmath84 and @xmath85 , and @xmath89 determined from the fits described in section iv.a .
.parameters for determining the potential eq.([eq : yukawa - n ] ) using the pomeron - quark coupling model defined by eqs.([eq : dl - jpsi-2])-([eq : alpha ] ) .
the predicted binding energies ( b.e . ) for proton ( @xmath203 , @xmath1 ( a=3 ) and @xmath204 ( a=12 ) are also listed . [ cols="^,^,^,^,^,^,^ " , ] we follow the procedure of subsection iv.b to assume that the @xmath162-@xmath0 bound states ( @xmath7_{j/\psi}$ ] ) are also defined by a potential , of yukawa form @xmath209 we expect that if a @xmath7_{j/\psi}$ ] bound state can be produced , its size must be small for color field to give strong attractive force .
thus it is reasonable to assume that the mean radius of @xmath7_{j/\psi}$ ] is close to the value @xmath210 of the initial @xmath162-@xmath15 system listed in table [ all ] .
we find that such a small size can be generated from choosing @xmath211 gev in defining the potential eq.([eq : yukawa - jpsi ] ) .
once a value of @xmath212 is chosen , we then determine the potential strength @xmath213 by requiring @xmath214 where @xmath215 is the @xmath0-@xmath162 relative wavefunction generated from the potential eq.([eq : yukawa - jpsi ] ) , and the values of @xmath210 for various considered cases are listed in table [ all ] .
the resulting @xmath213 and the binding energies ( b.e . )
are also listed there . here
we note that the binding energy increases as the mean radius @xmath210 and the corresponding probability @xmath216 decrease . with the wavefunctions for @xmath162-@xmath15 and @xmath162-@xmath0 specified in the previous subsections
, we can use the formula in section iii , with trivial changes in notations and spin quantum numbers , to calculate the total cross section of @xmath217_{j/\psi } + n$ ] .
however , we need to multiply the results by the probability @xmath216 of the @xmath15-@xmath7 $ ] component in @xmath1 ; namely the results from using eq.([eq : d - prod-1 ] ) is changed to @xmath218 \times [ \frac{d\sigma}{d\omega}]_0\ , , \label{eq : crst - q6}\end{aligned}\ ] ] where @xmath219_0 $ ] is calculated from using eq.([eq : d - prod-1 ] ) and all subsequent equations in section iii.a .
we first consider the case that @xmath220}=0.7\%$ ] as determined in refs.@xcite from fitting the @xmath8 phase shifts up to 1 gev . by using the parameters for models @xmath221 and @xmath222 listed in table [ all ]
, we obtain the results shown in fig.[fg:6q - a1a2 ] .
we observe that with the same small radius @xmath223 fm for the produced @xmath7_{j/\psi}$ ] system , the predicted cross sections are very close despite their potential range , measured by @xmath224 , and coupling constant @xmath213 can be very different .
the same finding is also from comparing the predicted cross sections from the models @xmath225 and @xmath226 , and also the models @xmath227 and @xmath228 . _
{ j/\psi } + n$ ] as function of the @xmath100-@xmath1 invariant mass @xmath46 .
the black solid and red dashed curves are for the case @xmath221 and @xmath222 in the table [ all ] , respectively . ] in the left - hand side of fig.[fg : diffabc ] , we show the dependence of the predicted cross sections on @xmath216 by comparing the cross sections from three models @xmath221 , @xmath225 , and @xmath227 listed in table [ all ] .
we observe that as @xmath229 decreases , the peak is shifted to higher energies .
each case has different threshold energy due to their differences in binding energies , as seen in table [ all ] .
their magnitudes are comparable despite their @xmath216 are very different .
we find that this is due to the fact that the cross section @xmath219_0 $ ] in eq.([eq : crst - q6 ] ) for the model with smaller @xmath230 is a factor of about 10 larger than that for the model with larger @xmath231 , since this large momentum transfer reaction favors the production of @xmath7_{j/\psi}$ ] with smaller size characterized by @xmath210 in table [ all ] .
the situation is similar to what we discussed in explaining the results shown in fig.[fig : compare ] .
thus the magnitudes of the cross sections from three models at peak positions are comparable because the factor of about 10 difference in @xmath219_0 $ ] in eq.([eq : crst - q6 ] ) is compensated by the similar factor of about 10 in @xmath229 .
however , the three models have rather different energy dependence , as also seen in the left - hand side of fig.[fg : diffabc ] .
on the other hand , they are all forward peaked , as shown in the right - hand side of fig.[fg : diffabc ] for the differential cross sections at @xmath232 gev .
the results shown in fig.[fg : diffabc ] suggest that the upper bound of the predicted total cross sections of @xmath9_{j/\psi } + n$ ] is about 2 - 4 pico - barn , @xmath225 , and @xmath227 of table [ all ] , respectively .
left : the total cross section of @xmath217_{j/\psi } + n$ ] as function of the @xmath100-@xmath1 invariant mass @xmath46 .
right : the differential cross section of @xmath233 } + n$ ] of the outgoing @xmath15 at @xmath100-@xmath1 invariant mass @xmath234 gev .
, title="fig : " ] , @xmath225 , and @xmath227 of table [ all ] , respectively . left : the total cross section of @xmath217_{j/\psi } + n$ ] as function of the @xmath100-@xmath1 invariant mass @xmath46 .
right : the differential cross section of @xmath233 } + n$ ] of the outgoing @xmath15 at @xmath100-@xmath1 invariant mass @xmath234 gev .
, title="fig : " ]
we have presented predictions of the cross sections of @xmath235_{j/\psi}$ ] reaction at energies near the @xmath0 production threshold . in the impulse approximation ,
the calculations have been performed by using several @xmath4 models based on the pomeron - exchange and pion - exchange mechanisms .
the @xmath0 wavefunctions in @xmath2_{j/\psi}$ ] are generated from various @xmath0-nucleus potentials which are constructed by either using a procedure based on the pomeron - quark coupling mechanism@xcite , or folding a @xmath0-n potential @xmath5 into the nuclear densities .
we consider @xmath5 derived from the effective field theory approach , lattice qcd , and pomeron - quark coupling model .
the upper bound of the predicted total cross sections is about 0.1 - 0.3 pico - barn
. clearly , our investigations are only for estimating the cross sections to facilitate the experimental considerations of possible measurements of @xmath2_{j/\psi}$ ] bound states at jefferson laboratory .
several improvements are needed for more quantitative predictions .
first we need precise data of @xmath10 near threshold to distinguish several models we have considered and also to develop a more sophisticated model.we also need the data to pin down the @xmath0-n interaction for a more realistic calculation of @xmath0-nucleus potential such as the folding model considered in this work .
one possibility is to use the @xmath236 reaction to extract the @xmath0-@xmath15 scattering length , as suggested in ref.@xcite .
alternatively , we can apply the model presented in this paper to determine the @xmath0-@xmath15 interactions by investigating the @xmath237 reaction .
possible experiments on these two processes will be very useful .
we of course also need to use more realistic wavefunctions for @xmath1 and @xmath3 while the s - wave oscillator wavefunctions employed in this investigation are reasonably consistent with the charge form factors calculated from the conventional nuclear models . motivated by the previous investigations@xcite on the effects due to multi - quark clusters in @xmath28 and @xmath11 , we have also considered the possibility of the production of a @xmath7-j/\psi$ ] bound state due to a six - quark @xmath7 $ ] cluster in @xmath1 . the compound bag model of @xmath8 scattering and the quark cluster model of nuclei
are used to estimate the @xmath7$]-n wavefunction in @xmath1 by imposing the condition that the sum of the contributions from @xmath7$]-n and @xmath238 components to the @xmath1 charge form factor must be consistent with what are predicted by the conventional nuclear models@xcite which explain the data very well .
the upper bound of the predicted total cross sections of @xmath9_{j/\psi } + n$ ] is about 2 - 4 pico - barn , depending on the model of @xmath4 used in the calculations .
if such bound states can be identified , it will open up a new window for investigating the role of the gluon field in determining the hadron structure .
we thank kawtar hafidi for the discussions on the possible @xmath0 production experiments at jeferson laboratory , henning esbensen and rocco schiavilla for their help in our bound state calculations .
this work is supported by the u.s .
department of energy , office of nuclear physics division , under contract no .
de - ac02 - 06ch11357 .
this research used resources of the national energy research scientific computing center , which is supported by the office of science of the u.s .
department of energy under contract no .
de - ac02 - 05ch11231 , and resources provided on `` fusion , '' a 320-node computing cluster operated by the laboratory computing resource center at argonne national laboratory .
jaffe and f. low , phys .
d * 19 * , 2105 ( 1979 ) ; f. low in * pointlike structure inside and outside hadrons * , proceedings of 1979 erice summer school , editted by a. zichichi ( plenum , new york , 1979 ) , p. 155 .
p. v. landshoff and o. nachtmann , z. phys .
c * 35 * , 405 ( 1987 ) .
laget and r. mendez - galain , nucl .
a581 * , 397 ( 1995 ) .
m. a. pichowsky and t .- s .
h. lee , phys .
d * 56 * , 1644 ( 1997 ) .
a. i. titov , t .- s .
h. lee , phys .
c * 67 * , 065205 ( 2003 ) .
m. e. binkley , c. bohler , j. butler , j. p. cumalat , i. gaines , m. gormley , d. harding and r. l. loveless _ et al .
_ , phys .
* 48 * , 73 ( 1982 ) .
b. h. denby , v. k. bharadwaj , d. j. summers , a. m. eisner , r. g. kennett , a. lu , r. j. morrison and m. s. witherell _ et al .
_ , phys .
* 52 * , 795 ( 1984 ) .
r. barate _
[ na14 collaboration ] , z. phys .
c * 33 * , 505 ( 1987 ) .
l. frabetti _ et al .
_ [ e687 collaboration ] , phys . lett .
b * 316 * , 197 ( 1993 ) .
u. camerini , j. g. learned , r. prepost , c. m. spencer , d. e. wiser , w. ash , r. l. anderson and d. ritson _ et al .
_ , phys .
lett . *
35 * , 483 ( 1975 ) .
b. gittelman , k. m. hanson , d. larson , e. loh , a. silverman and g. theodosiou , phys .
lett . * 35 * , 1616 ( 1975 ) .
r. l. anderson , excess muons and new results in @xmath239 photoproduction .
slac - pub-1471 ( unpublished ) .
s. aid _ et al .
_ [ h1 collaboration ] , nucl .
b * 472 * , 3 ( 1996 ) .
a. aktas _ et al .
_ [ h1 collaboration ] , eur .
j. c * 46 * , 585 ( 2006 ) .
j. breitweg
_ et al . _
[ zeus collaboration ] , z. phys .
c * 76 * , 599 ( 1997 ) .
s. chekanov _ et al . _
[ zeus collaboration ] , eur .
j. c * 24 * , 345 ( 2002 ) .
s. chekanov _ et al .
_ [ zeus collaboration ] , nucl .
b * 695 * , 3 ( 2004 ) .
m. derrick _ et al . _
[ zeus collaboration ] , phys .
b * 350 * , 120 ( 1995 ) . | the photo - production of @xmath0-@xmath1 bound state ( @xmath2_{j/\psi}$ ] ) on a @xmath3 target has been investigated using the impulse approximation .
the calculations have been performed using several @xmath4 models based on the pomeron - exchange and accounting for the pion - exchange mechanism at low energies .
the @xmath0 wavefunctions in @xmath2_{j/\psi}$ ] are generated from various @xmath0-nucleus potentials which are constructed by either using a procedure based on the pomeron - quark coupling mechanism or folding a @xmath0-n potential ( @xmath5 ) into the nuclear densities .
we consider @xmath5 derived from the effective field theory approach , lattice qcd , and pomeron - quark coupling mechanism .
the upper bound of the predicted total cross sections is about @xmath6 pico - barn .
we also consider the possibility of photo - production of a six quark-@xmath0 bound state ( @xmath7_{j/\psi})$ ] on the @xmath1 target .
the compound bag model of @xmath8 scattering and the quark cluster model of nuclei are used to estimate the @xmath7$]-n wavefunction in @xmath1 by imposing the condition that the calculated @xmath1 charge form factor must be consistent with what is predicted by the conventional nuclear model .
the upper bound of the predicted total cross sections of @xmath9_{j/\psi } + n$ ] is about 2 - 4 pico - barn , depending on the model of @xmath4 used in the calculations .
our results call for the need of precise measurements of @xmath10 and also the @xmath11 reactions near the threshold . |
massive stars in the range of 8 to @xmath3 130@xmath4 undergo core - collapse at the end of their evolution and become type ii and ib / c supernovae unless the entire star collapses into a black hole with no mass ejection .
these type ii and ib / c supernovae ( as well as type ia supernovae ) release large explosion energies and eject explosive nucleosynthesis materials , thus having strong dynamical , thermal , and chemical influences on the evolution of interstellar matter and galaxies .
therefore , the explosion energies of core - collapse supernovae are fundamentally important quantities , and an estimate of @xmath5 ergs has often been used in calculating nucleosynthesis and the impact on the interstellar medium .
( in the present paper , we use the explosion energy @xmath6 for the final _ isotropic _ kinetic energy of explosion . ) a good example is sn1987a in the large magellanic cloud , whose energy is estimated to be @xmath7 - 1.5 ) @xmath8 @xmath9 ergs from its early light curve @xcite .
one of the most interesting recent developments in the study of supernovae ( sne ) is the discovery of the connection between long - duration gamma ray bursts ( grbs ) and a particular subtype of core - collapse sne ( type ic ) as has been clearly established from grb 980425/sn 1998bw @xcite , grb 030329/sn 2003dh @xcite , and grb 031203/sn 2003lw @xcite .
these grb - sne have similar properties ; they are all hypernovae , i.e. , very energetic supernovae , whose _ isotropic _ kinetic energy ( ke ) exceeds @xmath10erg , about 10 times the ke of normal core - collapse sne ( hereafter @xmath11erg ) .
figure 1 shows the spectrum of sn 1998bw / grb 980425 a few days before maximum .
the spectrum of sn 1998bw has very broad lines , indicative of the large @xmath6 .
the strongest absorptions are ti ii - fe ii(shortwards of @xmath12 , fe ii - fe iii ( near 4500 ) , si ii(near 5700 ) , o i - ca ii ( between 7000 and 8000 ) . from the synthetic spectra and light curves , it was interpreted as the explosion of a massive star , with @xmath13erg and @xmath14.@xcite also the very high luminosity of sn 1998bw indicates that a large amount of @xmath15ni ( @xmath16 ) was synthesized in the explosion .
the other two grb - sne , 2003dh and 2003lw , are also characterized by the very broad line features and the very high luminosity . @xmath17 and @xmath18 are estimated from synthetic spectra and light curves and summarized in figure 2@xcite .
it is clearly seen that grb - sne are the explosions of massive progenitor stars ( @xmath19 ) , have large explosion kinetic energies ( @xmath20erg ) , synthesized large amounts of @xmath15ni ( @xmath21 ) , thus forming the `` hypernova branch '' in figure 2 .
other `` non - grb hypernovae '' , such as sn 1997ef @xcite , sn 2002ap @xcite , and sn 2003jd @xcite , have been observed .
these hne show spectral features similar to those of the grb - sne but are not known to have been accompanied by a grb .
the estimated @xmath17 and @xmath6 , obtained from synthetic light curves and spectra , show that there is a tendency for non - grb hne to have smaller @xmath17 , @xmath6 , and lower luminosities as summarized in figure 2 .
for example , sn 2002ap has similar spectral features , but narrower and redder ( fig .
1 ) , which was modeled as a smaller energy explosion , with @xmath22erg and @xmath23@xcite .
recently x - ray flash ( xrf ) 060218 has been found to be connected to sn ic 2006aj @xcite .
sn 2006aj is very similar to sn 2002ap , being a less energetic ( @xmath24 ) .
however , it differs in the lack of the o i - ca ii absorption near 7500 .
the progenitor is estimated to be @xmath25 , thus being suggested to be a `` neutron star - making sn '' @xcite ( see 2 for details ) .
+ hypernovae are also characterized by asphericity from the observations of polarization and emission line features ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
the explosion energy of the aspherical models for hypernovae tends to be smaller than the spherical models by a factor of 2 - 3 , but still being as high as @xmath26 @xcite .
in contrast to hne , sne ii 1997d and 1999br were very faint sne with very low ke @xcite .
in the diagram that shows @xmath6 and the mass of @xmath15ni ejected @xmath27ni ) as a function of the main - sequence mass @xmath28 of the progenitor star ( figure [ figme ] ) , therefore , we propose that sne from stars with @xmath29 have different @xmath6 and @xmath27ni ) , with a bright , energetic `` hypernova branch '' at one extreme and a faint , low - energy sn branch at the other @xcite . for the faint sne ,
the explosion energy was so small that most @xmath15ni fell back onto the compact remnant ( e.g. , * ? ? ? * ) .
thus the faint sn branch may become a `` failed '' sn branch at larger @xmath28 . between the two branches , there may be a variety of sne @xcite .
this trend might be interpreted as follows .
stars more massive than @xmath3 25 @xmath4 form a black hole at the end of their evolution .
stars with non - rotating black holes are likely to collapse `` quietly '' ejecting a small amount of heavy elements ( faint supernovae ) . in contrast , stars with rotating black holes are likely to give rise to hypernovae .
the hypernova progenitors might form the rapidly rotating cores by spiraling - in of a companion star in a binary system .
here we focus on our findings that sn ic 2006aj ( 2 ) and sn ib 2005bf ( 3 ) are very different sne from previously known sne / hne .
these properties might be due to their progenitor masses , which indicate that these sne correspond to the border from the ns and bh formation . as a related topic ,
it is of vital importance to identify the first generation stars in the universe , i.e. , totally metal - free , pop iii stars .
we examine possible grb / hypernova - first star connection through nucleosnthesis approach @xcite .
we summarize nucleosynthesis features in hypernovae , which must show some important differences from normal supernova explosions ( 4 ) .
this might be related to the unpredicted abundance patterns observed in the extremely metal - poor halo stars .
this is one of the important challenges of the current astronomy @xcite .
grb060218 is located in a galaxy only @xmath30mpc away and it is the second closest event as ever .
the grb was weak @xcite , as is often the case for nearby ones @xcite , and was classified as x - ray flash ( xrf ) because of its soft spectrum .
as the grb was not followed by a bright afterglow , the presence of a sn 2006aj was soon confirmed @xcite .
here we summarize the properties of sn 2006aj by comparing with other sne ic .
sn 2006aj has several features that make it unique .
it is less bright than the other grb / sne ( figure [ fig06ajlc ] ) .
its rapid photometric evolution is very similar to that of a dimmer , non - grb sn 2002ap @xcite , but it is somewhat faster . although its spectrum is characterized by broader absorption lines as in sn 1998bw and other grb / sn , they are not as broad as those of sn 1998bw , and again it is much more similar to that of sn 2002ap ( figure [ fig06ajsp ] ) .
the most interesting property of sn 2006aj is surprisingly weak oxygen lines , much weaker than in type ic sne . in order to quantify its properties , we modeled the spectra of sn 2006aj with a radiation transfer code as in @xcite .
we first employed the same explosion model applied for sn 2002ap @xcite .
the model has an ejected mass @xmath17 @xmath31 and a kinetic energy @xmath22erg .
this explosion model gives reasonable fits to the spectra .
however , in order to improve the match , we had to reduce the masses of both oxygen and calcium substantially and reduce @xmath17 accordingly ( figure [ figspfit ] ) . as a result ,
we derive for sn 2006aj @xmath32 and @xmath33erg .
lack of oxygen in the spectra does not necessarily mean absence of oxygen in the ejecta .
our model contains @xmath34 of o , and oxygen is still the dominant element .
the strength of the oi@xmath357774 line , which is the strongest oxygen line in optical wavelength , is sensitive to the temperature in the ejecta .
since the fraction of oi is larger in the lower temperature ejecta ( although oii is still the dominant ionization state ) , the normal sne ib / c always show the strong oi absorption ( see sn 1994i in fig [ fig06ajsp ] ) irrespective of the ejecta mass . in more luminous sne like grb - sne and sn 2006aj
, the oi fraction tends to be smaller .
however , if the ejecta are very massive , e.g. , @xmath36 , the mass of oi is large enough to make the strong absorption ( see sn 1998bw in fig [ fig06ajsp ] ) . in the case of sn 2006aj , the temperature is larger than in normal sne ib / c .
therefore , the weak oi line indicates that the ejecta mass is not as massive as sn 1998bw , which supports our conclusion .
the spectroscopic results are confirmed through the light curve modeling .
the timescale of the lc around maximum brightness reflects the timescale for optical photons to diffuse @xcite .
for the more massive ejecta and the smaller kinetic energy , the lc peaks later and the lc width becomes broader becaause it is more difficult for photons to escape .
we synthesize the theoretical light curve with the 1-dimensional density and chemical abundance structure that we find in the above spectroscopic analysis .
we then compare it with the optical - infrared bolometric light curve of sn 2006aj .
the best match to the rapidly rising light curve is achieved with a total @xmath15ni mass of @xmath37 in which @xmath38 is located above 20,000 km s@xmath39(figure [ figlcfit ] ) .
the high - velocity @xmath15ni is responsible for the fast rise of the light curve , because photons created can escape more easily . in the model ,
the mass fraction of @xmath15ni in the high velocity region is as large as @xmath3 35% , which is unlikely to be attained in a spherically symmetric explosion . in a realistic asymmetric explosion
, the high - velocity @xmath15ni could abundantly be produced along the direction of the grb jets @xcite .
the properties of the sn 2006aj ( smaller energy , smaller ejected mass ) suggest that sn 2006aj is not the same type of event as the other grb - sne known thus far .
one possibility is that the initial mass of the progenitor star is much smaller than the other grb - sne , so that the collapse / explosion generated less energy .
if the zero - age main sequence mass is @xmath40 , for example , the star would be at the boundary between collapse to a black hole or to a neutron star . in this mass range
, there are indications of a spread in both @xmath6 and the mass of @xmath15ni synthesized @xcite . the fact that a relatively large amount of @xmath15ni is required in sn
2006aj possibly suggests that the star collapsed only to a neutron star because more core material would be available to synthesize @xmath15ni in the case
. although the kinetic energy of @xmath41 erg is larger than the canonical value ( @xmath42 erg , @xcite ) in the mass range of @xmath43 , such an energy may be easily attainable .
additionally , magnetar - type activity may have been present , increasing the explosion energy@xcite .
it is conceivable that in this weaker explosion than typical grb - sne , the fraction of energy channeled to relativistic ejecta is smaller , giving rise to an xrf rather than a classical grb .
another case of a sn associated with an xrf has been reported ( xrf030723)@xcite . the putative sn , although its spectrum was not observed , was best consistent with the properties of sn 2002ap @xcite .
this may suggest that xrfs are associated with less massive progenitor stars than those of canonical grbs , and that the two groups may be differentiated by the formation of a neutron star @xcite or a bh .
still , the progenitor star must have been thoroughly stripped of its h and he envelopes , which is a general property of all grb - sne and probably a requirement for the emission of a high energy transient .
these facts may indicate that the progenitor is in a binary system .
if magnetars are related to the explosion mechanism , some short @xmath44-ray repeaters energized by a magnetar @xcite may be remnants of grb060218-like events .
magnetars could generate a grb at two distinct times .
as they are born , when they have a very large spin rate ( @xmath45 ms ) , an xrf ( or a soft grb ) is produced as in sn2006aj / grb060218 .
later ( more than 1,000 yrs ) , when their spin rate is much slower , they could produce short - hard grbs @xcite .
stars of mass @xmath46 are much more common than stars of @xmath47 , and so it is highly likely that events such as grb060218 are much more common in nature than the highly energetic grbs .
they are , however , much more difficult to detect because they have a low @xmath44-ray flux .
the discovery of grb060218/sn 2006aj suggests that there may be a wide range of properties of both the sn and the grb in particular in this mass range .
the continuing study of these intriguing events will be exciting and rewarding .
peculiar and inhomogeneous natures of supernovae originating from @xmath48 , i.e. , boundary between a neutron star formation and a black hole formation , are further highlighted by sn 2005bf .
unlike sn 2006aj / grb060218 , it did not show a high energy transient counter part .
however , it did show very unique photometric and spectroscopic behavior .
based on the calculated light curve and the spectra , we believe that sn2005bf fits into the scheme suggested by @xcite which places core - collapse sn in a sequence ( iip - iil - iib - ib - ic ) of increasing mass loss from the progenitor star , and that it originated in collapse of a star with @xmath49 .
sn 2005bf was discovered by @xcite on april 6 , 2005 ( ut ) in the spiral arms of the sbb galaxy mcg + 00 - 27 - 5 .
it was initially classified as a type ic sn ( sn ic ) @xcite . as time went by ,
he lines were increasingly developed .
then it was classified as type ib @xcite .
even stranger , the light curve is very different from any known sn @xcite : a fairly rapid rise to a first peak was followed by a period of stalling or slow decline and by a new rise to a later , brighter peak at @xmath50 days after explosion ( fig .
[ fig05bflc ] ) .
the brightness ( @xmath51 mag ) at the relatively slow peak date suggests that a large amount of @xmath15niis ejected .
sn 2005bf does not show the broad lines seen in hypernovae .
these properties make sn 2005bf a very interesting sn .
to derive physical quantities from observed properties , we first try fitting the light curve covering the first 80 days @xcite .
the radioactive decay ( @xmath15ni @xmath52 @xmath15co @xmath52@xmath15fe ) is responsible for powering the light curve .
the theoretical lc width near peak depends on ejected mass @xmath17 and explosion kinetic energy @xmath6 as @xmath53 @xcite .
the mass and distribution of @xmath15ni are constrained by the lc brightness and shape .
in general , various combinations of ( @xmath17 , @xmath6 ) can fit the lc .
the degeneracy can be solved by performing spectral modeling , since the line width is scaled as @xmath54 .
we compute a set of synthetic light curves for a he star model , by varying ( @xmath17 , @xmath6 ) and abundance distribution ( including distribution of the heating source @xmath15ni ) .
parameters are constrained by the observed bolometric light curve ( lc ) constructed as in @xcite ( see fig .
[ fig05bflc ] ) . we assumed a galactic reddening @xmath55 , a distance modulus @xmath56 , and an explosion date of @xmath57 march 28 ut as inferred from the marginal detection on 2005 march 30 ut @xcite .
the light curve for a model with @xmath58 and @xmath59 is shown as a dashed line in figure [ fig05bflc ] .
the model curve yields a nice fit until the second , main peak is reached . however , the observed lc declines rapidly thereafter , unlike other well - observed sne ib / c . a possible solution for this is to consider a situation in which the ejecta are more transparent to gamma - rays than in other sne ib / c ( usually @xmath60 @xcite ) . using @xmath61 at @xmath62 km s@xmath39 the light curve shown in figure [ fig05bflc ] is obtained .
the model parameters are the following : @xmath58 , @xmath63 , @xmath64 .
the parameters are not so different from those derived without accelerated gamma - ray escape , since the set of ( @xmath17 , @xmath6 ) is constrained by the diffusion time scale and expansion time scale , and @xmath65(@xmath15ni ) is determined from the peak brightness .
these are basically independent from the lc behavior after the peak date . to break the degeneracy in ( @xmath17 , @xmath6 ) and select the most likely model , synthetic spectra are computed and compared to the observed ones in figure [ fig05bfsp ] .
the model with ( @xmath66 , @xmath67 ) = ( 7 , 1.3 ) provides satisfactory fits for all spectra . at the time of the first peak ( ut april 13
, 16 days after explosion ; @xcite ) sn 2005bf exhibited sn ic features , but actually both he and h can be seen as weak features .
the feature near 5700 is probably he i 5876 . a photospheric velocity of @xmath68 km s@xmath39is obtained by the spectroscopic model .
the model with @xmath58 fits better than that with @xmath69 .
the feature at 6300 is partly attributed to h@xmath70 ( fig .
[ fig05bfsp ] , inset ) , indicating presence of @xmath71 of hydrogen above @xmath72 km s@xmath39 .
the spectra near maximum brightness ( ut may 4 , 37 days after explosion ) and on ut may 16th ( 49 days ) are reasonably reproduced by the same ejecta model , with the photospheric velocity @xmath73 km s@xmath39 and @xmath74 km s@xmath39 , respectively . according to the model , the ejecta properties
are derived as follows : @xmath75 and @xmath76 .
the ejecta consist of @xmath15ni ( @xmath77 ) mostly near the center , he ( @xmath78 ) , intermediate mass elements ( mainly o , si , s ) , and a small amount of h ( @xmath71 ) . thus the progenitor had lost almost all its h envelope , but retained most of the he - rich layer ( a wn star ) .
the he core mass at the explosion was @xmath79 .
the progenitor was probably a wn star of main - sequence mass @xmath80 @xcite .
the formation of a wn star from a star of only @xmath81 suggests that rotation may have been important @xcite , although not sufficient to make sn 2005bf a hypernova . in order to produce @xmath82 @xmath15ni , the mass cut that separates the ejecta and
the compact remnant should be as deep as @xmath83 .
this suggests that the remnant was a neutron star rather than a black - hole .
the mass range @xmath84 is near the transition from neutron star ( sn 2005bf ) to black hole formation ( sn 2002ap ; @xcite ) , the exact boundary depending on rotation and mass loss .
the ejecta we derived have small @xmath6 relative to @xmath17 . in this way
, relatively delayed peak at @xmath50 days is explined .
the rapid rise to the first peak at @xmath85 days requires a small amount of @xmath15ni ( @xmath86 ) at high velocity ( @xmath87 km s@xmath39 ) , while most of @xmath15ni is neat the center below @xmath88 km s@xmath39 .
the high velocity component of @xmath15ni is at the bottom of the he layer .
it may express @xmath15ni - rich jets or blobs that did not reach the he layer .
such inhomogeneous structure might result in the enhanced @xmath44-ray escape as is assumed to obtain a better fit to the lc after the second peak . at late epochs
( @xmath45 year since the explosion ) , a sn enters into nebular phases .
then a sn shows optically thin spectra dominated by emission lines arising even from the deepest regions .
thus the nebular phase observations can prove the inner regions of the ejecta , which are not seen at early phases .
spectroscopy and photometry of sn 2005bf have been performed on 2005 december 26 ( ut ) and on 2006 february 6 with the 8.2 m subaru telescope equipped with the faint object camera and spectrograph ( focas ; @xcite ) . detailed analysis for the late time observation
is presented elsewhere @xcite .
figure [ figneb ] shows the reduced spectra of sn 2005bf on 2005 december 26 ( @xmath89days since the putative explosion date ) . there is a feature at @xmath90 with fwhm @xmath91 km s@xmath39
this is most likely h@xmath92 emission .
this supports the existence of the thin h envelope described in [ sec : bf - early ] the spectrum shows strong forbidden emission lines , in which [ oi ] @xmath936300 , 6363 and [ caii ] @xmath357300 are the strongest .
other weak emission features are visible at @xmath94 ( [ feii ] ) and @xmath95(caii ir and [ ci ] @xmath358727 ) .
the [ oi ] @xmath356300/[caii ] @xmath357300 ratio in sn 2005bf is relatively small , indicating the progenitor star with @xmath96 . a forbidden carbon line [ ci ] @xmath358727 looks strong , indicating high c / o ratio ( @xmath97 ) in the ejecta .
this suggests the progenitor star with @xmath98 . from these analyses ,
the nebular spectra supports the progenitor mass @xmath49 , as derived in the earlier phases ( [ sec : bf - early ] ) .
profiles of these line profiles are shown in figure [ figneb ] .
interestingly , all these lines show blueshift relative to the rest wavelength ( @xmath99 km s@xmath39 ) .
a simple and straightforward interpretation is that we see a unipolar explosion , on average moving toward us , as is expected from the fast - moving @xmath15ni seen in the early phase ( [ sec : bf - early ] ) .
another interpretation is self - absorption of the light within the ejecta which reduces the light from the far ( therefore red ) side .
these are thoroughly discussed in maeda et al .
( 2007 ) @xcite .
the major features of sn 2005bf can be understood in the context of the explosion of a wn star with the progenitor @xmath100 .
however , there are still some questions to be answered , which are probably related to unique natures of the progenitor , the central remnant , and the explosion physics of stars with @xmath101 . first , in early phases , he lines evolved in their strengths and velocities in a unique way .
they become stronger ( therefore look like sn ic first , then like sn ib afterward ) as time goes by .
their velocities are also increased as a function of time . at the first peak , these lines are well explained in lte level populations . near maximum brightness ( ut may 4 ,
37 days after explosion ) and on ut may 16th ( 49 days ) , however , level populations of he ion should be more abundant than in lte populations by a factor of @xmath102 ( at @xmath103 km s@xmath39 ) and @xmath104 ( at @xmath105 km
s@xmath39 ) , respectively .
non - thermal effects resulting from radioactive decay gamma - rays are believed to be essential to populate hei levels @xcite .
why sn 2005bf showed the peculiar evolution of he lines is probably related to its unique @xmath15ni distribution .
this is still an open question , which will probably provide further clue to understand the natures of sn 2005bf .
another question was brought by the nebular phase observation .
the r - band magnitude at 2006 december 26 ( @xmath89 days since the explosion ) is @xmath106 , corresponding the absolute magnitude @xmath107 after correcting the distance and the reddening .
it is very faint as compared to other sne ib / c , at least by 2 magnitudes ( e.g. , by 3 magnitudes fainter than sn 1998bw at a similar epoch ) . if the r magnitude is close the bolometric magnitude ( which is usually a good approximation for sn ib / c nebulae if most of the light is emitted in optical ranges ) , @xmath15ni required to fit the luminosity is only @xmath108 .
why this is much smaller than that reproducing the early phase peak luminosity ( at @xmath50 days ) is still to be answered .
several possibilities , including late time fallback of materials onto a central remnant and a magnetar - like activity are worth studying @xcite .
in core - collapse supernovae / hypernovae , stellar material undergoes shock heating and subsequent explosive nucleosynthesis .
iron - peak elements are produced in two distinct regions , which are characterized by the peak temperature , @xmath109 , of the shocked material . for @xmath110k
, material undergoes complete si burning whose products include co , zn , v , and some cr after radioactive decays . for @xmath111k @xmath112k ,
incomplete si burning takes place and its decayed products include cr and mn ( e.g. , @xcite ) .
the right panel of figure [ fig09 ] shows the composition in the ejecta of a 25 @xmath4 hypernova model ( @xmath113 ) .
the nucleosynthesis in a normal 25 @xmath4 sn model ( @xmath114 ) is also shown for comparison in the left panel of figure [ fig09 ] @xcite .
we note the following characteristics of nucleosynthesis with very large explosion energies @xcite : \(1 ) both complete and incomplete si - burning regions shift outward in mass compared with normal supernovae , so that the mass ratio between the complete and incomplete si - burning regions becomes larger . as a result ,
higher energy explosions tend to produce larger [ ( zn , co , v)/fe ] and smaller [ ( mn , cr)/fe ] , which can explain the trend observed in extremely metal - poor stars @xcite . ( here [ a / b ] @xmath115 , where the subscript @xmath116 refers to the solar value and @xmath117 and @xmath118 are the abundances of elements a and b , respectively . )
\(2 ) in the complete si - burning region of hypernovae , elements produced by @xmath70-rich freezeout are enhanced .
hence , isotopes synthesized through capturing of @xmath70-particles , such as @xmath119ti , @xmath120cr , and @xmath121ge ( decaying into @xmath119ca , @xmath120ti , and @xmath121zn , respectively ) are more abundant .
\(3 ) oxygen burning takes place in more extended regions for the larger ke .
then more o , c , al are burned to produce a larger amount of burning products such as si , s , and ar .
therefore , hypernova nucleosynthesis is characterized by large abundance ratios of [ si , s / o ] , which can explain the abundance feature of m82 @xcite . hypernova nucleosynthesis may have made an important contribution to galactic chemical evolution . in the early galactic epoch when the galaxy was not yet chemically well - mixed , [ fe / h ] may well be determined by mostly a single sn event @xcite .
the formation of metal - poor stars is supposed to be driven by a supernova shock , so that [ fe / h ] is determined by the ejected fe mass and the amount of circumstellar hydrogen swept - up by the shock wave @xcite .
then , hypernovae with larger @xmath6 are likely to induce the formation of stars with smaller [ fe / h ] , because the mass of interstellar hydrogen swept up by a hypernova is roughly proportional to @xmath6 @xcite and the ratio of the ejected iron mass to @xmath6 is smaller for hypernovae than for normal supernovae .
the `` mixing and fall back '' process can reproduce the abundance pattern of the typical emp stars without a disagreement between [ ( light element)/fe ] and [ ( fe - peak element)/fe ] .
figure [ fig11 ] shows that the averaged abundances of [ fe / h ] @xmath122 stars in @xcite can be fitted well with the hypernova model of 20 @xmath4 and @xmath123 10 ( upper ) but not with the normal sn model of 15 @xmath4 and @xmath123 1 ( lower ) @xcite .
in the observed abundances of halo stars , there are significant differences between the abundance patterns in the iron - peak elements below and above [ fe / h]@xmath124 - @xmath125 ( figure 12 ) .
\(1 ) for [ fe / h]@xmath126 , the mean values of [ cr / fe ] and [ mn / fe ] decrease toward smaller metallicity , while [ co / fe ] increases @xcite .
\(2 ) [ zn / fe]@xmath127 for [ fe / h ] @xmath128 to @xmath129 @xcite , while at [ fe / h ] @xmath130 , [ zn / fe ] increases toward smaller metallicity @xcite .
the larger [ ( zn , co)/fe ] and smaller [ ( mn , cr)/fe ] in the supernova ejecta can be realized if the mass ratio between the complete si burning region and the incomplete si burning region is larger , or equivalently if deep material from the complete si - burning region is ejected by mixing or aspherical effects .
this can be realized if ( 1 ) the mass cut between the ejecta and the compact remnant is located at smaller @xmath131 @xcite , ( 2 ) @xmath6 is larger to move the outer edge of the complete si burning region to larger @xmath131 @xcite , or ( 3 ) asphericity in the explosion is larger @xcite . among these possibilities ,
a large explosion energy @xmath6 enhances @xmath70-rich freezeout , which results in an increase of the local mass fractions of zn and co , while cr and mn are not enhanced @xcite .
models with @xmath132 do not produce sufficiently large [ zn / fe ] . to be compatible with the observations of [ zn / fe ] @xmath133 , the explosion energy must be much larger , i.e. , @xmath26 for @xmath134 , i.e. , hypernova - like explosions of massive stars ( @xmath134 ) with @xmath135 are responsible for the production of zn .
figure 12 exhibits that the higher - energy models tend to be located at lower [ fe / h ] = log@xmath136 ( fe/@xmath137 ( if @xmath138 distribute around a certain peak value ) , and thus can explain the observed trend . in the hypernova models ,
the overproduction of fe , as found in the simple `` deep '' mass - cut model , can be avoided with the mixing - fallback model @xcite .
therefore , if hypernovae made significant contributions to the early galactic chemical evolution , it could explain the large zn and co abundances and the small mn and cr abundances observed in very metal - poor stars ( fig . [ fig11 ] : @xcite ) .
in contrast to the core - collapse supernovae of 20 - 130 @xmath4 stars , the observed abundance patterns can not be explained by the explosions of more massive , 130 - 300 @xmath4 stars .
these stars undergo pair - instability supernovae ( pisne ) and are disrupted completely ( e.g. , @xcite ) , which can not be consistent with the large c / fe observed in c - rich emp stars
. the abundance ratios of iron - peak elements ( [ zn / fe ] @xmath139 and [ co / fe ] @xmath140 ) in the pisn ejecta @xcite , can not explain the large zn / fe and co / fe in the typical emp stars @xcite and cs22949 - 037 @xcite either .
therefore the supernova progenitors that are responsible for the formation of emp stars are most likely in the range of @xmath141 @xmath4 , but not more massive than 130 @xmath4 .
this upper limit depends on the stability of massive stars .
we summarize the properties of core - collapse sne as a function of the progenitor mass .
as seen in figure 2 , three grb - sne are all similar hypernovae ( hne ) for their @xmath17 and @xmath6 . for non - grb hne , whether the non - dection of grbs is the effect of different orientations or of an intrinsic property is still a matter of debate , but there is a tendency for them to have smaller @xmath17 and @xmath6 .
the discovery of xrf 060218/sn 2006aj and their properties extend the grb - hn connection to xrfs and to the hn progenitor mass as low as @xmath142 .
the xrf 060218 may be driven by a neutron star rather than a black hole .
the progenitor mass range of 20 - 25 @xmath2 is particularly interesting , because it corresponds to the transition from the ns formation to the bh formation .
the nss from this mass range could be much more active than those from lower mass range because of possibly much larger ns masses ( near the maximum mass ) or possibly large magnetic field ( i.e. , magnetar ) .
possible xrfs and grbs from this mass range of 20 - 25 @xmath2 might form a different population . an estimate and a comparison for the rates of hypernovae and grbs was performed @xcite . within the substantial uncertainties ,
the estimates are shown to be quite comparable and give a galactic rate of @xmath143 to @xmath144 for both events .
these rates are several orders of magnitude lower than the rate of core - collapse supernovae , suggesting that the evolution leading to an hn / grb requires special circumstances , very likely due to binary interactions . the discovery of sn 2006aj / xrf 060218 affects the estimate of the grb rates .
the observations of various bands suggest xrf 060218 is an intrinsically weak and soft event , rather than a classical grb observed off - axis . the existence of a population of less luminous grb / xrfs than `` classical '' grbs is suggested .
such events may be the most abundant form of x- or gamma - ray explosive transient in the universe , but instrumental limits allow us to detect them only locally , so that several intrinsically sub - luminous bursts may remain undetected .
if the low - redshift grbs are really typical of the global grb population , then their discovery within the current time and sky coverage must be consistent with the local grb explosion rate as deduced from the very large batse grb sample .
pian et al.@xcite include this underluminous population , assume no correction for possible collimation , which may vary from object to object , and obtain a local grb rate of @xmath145 , compared to @xmath146 estimated from the cosmological events only .
the local rate of events that give rise to grbs is therefore at least one hundred times the rate estimated from the cosmological events only ( i.e. , those observed by batse ) .
the fraction of supernovae that are associated with grbs or xrfs may be higher than currently thought . | the connection between long gamma ray bursts ( grbs ) and supernovae ( sne ) , have been established through the well observed cases of grb980425/sn1998bw , grb030329/sn2003dh and grb031203/sn2003lw .
these events can be explained as the prompt collapse to a black hole ( bh ) of the core of a massive star ( @xmath0 ) that had lost its outer hydrogen and helium envelopes .
all these sne exhibited strong oxygen lines , and their energies were much larger than those of typical sne , thus these sne are called hypernovae ( hne ) .
the case of sn2006aj / grb060218 appears different : the grb was weak and soft ( an x - ray flash , xrf ) ; the sn is dimmer and has very weak oxygen lines .
the explosion energy of sn2006aj was smaller , as was the ejected mass .
in our model , the progenitor star had a smaller mass than other grb / sne ( @xmath1 ) , suggesting that a neutron star ( ns ) rather than a black hole was formed .
if the nascent neutron star was strongly magnetized ( a so - called magnetar ) and rapidly spinning , it may launch a weak grb or an xrf .
the final fate of 20 - 30 @xmath2 stars show interesting variety , as seen in the very peculiar type ib / c sn 2005bf .
this mass range corresponds to the ns to bh transition .
we also compare the nucleosynthesis feature of hne with the metal - poor stars and suggest the hypernova - first star connection .
supernovae ; hypernovae ; gamma - ray bursts ; first stars ; nucleosynthesis |
photonic crystals ( phcs ) , since first proposed by yablonovitch ( @xcite ) and john ( @xcite ) , have attracted considerable attention .
these artificial structures are materials patterned with a periodicity in dielectric constant , which can create a range of frequencies ( called a photonic band gap ) for which light is forbidden to exist within the interior of the crystal .
these photonic band gaps ( pbgs ) lend themselves to numerous applications in linear , nonlinear and quantum optics .
it has been predicted and confirmed experimentally that photonic crystals allow to modify and control spontaneous emission of an exited atom due to modification of the density of quantum states ( @xcite ) . in particular , it is well known , that suppression of spontaneous emission is important in keeping a system in exited states and can result in the reduction of noise in optoelectronic devices .
localized electromagnetic modes can be achieved by introduction of defects within the photonic crystal lattice ( @xcite ) .
the nature of the defect decides on the shapes and properties of localized photonic states in the gap .
a point defect could act like a microcavity , a line defect like a waveguide . in this paper , we study the effect of defects on the density of states and simultaneously spontaneous emission rate in one - dimensional photonic crystal .
we begin by presenting an analytical model of density of states which is called the effective resonator model ( section [ secmodel ] ) . in general
it is well known , that the microcavity affects density of states and spontaneous emission rate . in our approach effective cavity model
is used to study density of states and spontaneous emission rate in finite photonic crystal . moreover ,
the influence of local defects on density of states in spectral and spatial domain is considered .
presented model also takes into consideration direction of propagation and therefore does not neglect anisotropic nature of photonic crystals . in section [ secms ]
we briefly derive the analytical expression for mode spectrum which is related to the density of states .
this relation allow to analyze properties of one - dimensional photonic crystals with defects .
the influence of defects in 1d photonic crystal on the density of states and spontaneous emission , in spatial and frequency domain is investigated in section [ secdef ] .
conclusions can be found in section [ secconc ] .
it is an easy task to calculate reflection coefficient of a lattice built of dielectric layers , which extend towards infinity in @xmath0 and @xmath1 directions ( where coordinate system is taken as shown in fig .
[ pcstr ] ) . following @xcite , in every layer , for each of two polarizations , a solution of maxwell equations can be written as a superposition of two plane waves : @xmath2 where @xmath3 stands for the number of the layer and @xmath4 these plane waves are bound with plane waves in the next layers by continuity conditions , which can be written as a matrix equation ( @xcite ) : @xmath5 where @xmath6 is the continuity condition matrix : @xmath7\ , \ ] ] @xmath8 \ , \ ] ] where @xmath9 is the wave vector s @xmath10 component in the @xmath3th layer , @xmath11 is the @xmath3th layer s width and @xmath12 is the @xmath3th layer s refractive index ( more precisely , eq . is a relation between amplitudes of fields in different points of space left edges of layers ) .
amplitudes of plane waves outside the lattice are therefore related to each other : @xmath13 if @xmath14 is an amplitude of the plane wave incident on the lattice , then @xmath15 corresponds to the plane wave reflected in the infinity .
this reflection does not occur , hence @xmath16 .
therefore : @xmath17 where @xmath18\equiv m_n\ldots m_1m_0\ ] ] and as a result : @xmath19 the reflection coefficient of the whole lattice is then given by @xmath20 with the assumption , that amplitudes @xmath14 and @xmath21 describe incident and reflected plane waves in a `` virtual '' layer with the same refraction coefficient as the external medium and width @xmath22 .
the above expression is purely analytical , though for lattices with diversified layers it would take a very complicated form if written as an explicit function of wave vector , polarization and lattice parameters , which are hidden in @xmath23 and @xmath24 .
having derived , we can calculate reflection coefficient of any number of dielectric layers , what allows us to treat any randomly selected layer of a lattice as an effective resonator .
the idea of this model is presented in fig .
[ pctor ] .
formally , a lattice can be treated as two lattices and the layer joined together .
it is possible to replace the lattices on both sides of the layer with mirrors having the same reflection coefficients @xmath25 and @xmath26 , where @xmath27 denotes polarization .
let @xmath28 be a plane wave exciting a resonator limited by two parallel planes with reflection coefficients @xmath29 and @xmath30 ( direction of this vector determines polarization of the wave ) . because of reflections , there appears in the resonator an effective plane wave @xmath31 .
we introduce an operator @xmath32 , which establishes a relation between them ( @xcite ) : @xmath33 we call a plane wave @xmath34 with polarization @xmath27 and wave vector @xmath35 a mode of the resonator , if it is an eigenvector of @xmath32 : @xmath36 eigenvalue @xmath37 is proportional to mode spectrum @xmath38 , which we define as : @xmath39 mode spectrum describes distribution of modes in wave vector domain , which is equivalent to frequency and direction of propagation .
we consider the following relation between mode spectrum and density of states ( dos ) : @xmath40 in free space no reflections occur and therefore the operator @xmath41 is an identity , with all eigenvalues equal to @xmath42 .
hence , mode spectrum is constant in free space , and with taken normalization we obtain from valid and well known formula for dos in free space : @xmath43 let @xmath34 be a mode of the considered resonator . if the amplitude of the excitation is @xmath44 , to find the amplitude @xmath45 of the effective plane wave one has to sum all the reflected waves : @xmath46 where @xmath47 is the resonator s width .
the second sum arises because continuity conditions apply on both mirrors , and it can be easily checked , that for a plane wave incident on a plane between two different media the change of sign of @xmath48 ( assuming that @xmath10 axis is perpendicular to the plane ) causes the reflection coefficient to become its conjugate .
it follows , that mode spectrum for the resonator is described with the formula : @xmath49 where reflection coefficients depend on polarization and wave vector .
and number of elementary cell @xmath50 . ] formula allows to analyze properties of layers of a lattice , in a particular case a 1d photonic crystal .
in this kind of structure there appear photonic band gaps , regions of frequency for which no mode is allowed . in terms of mode spectrum a photonic band gap is a region , for which mode spectrum is equal to @xmath51 .
however , in a structure built of finite number of layers it is not strictly possible , because the structure is not able to fully eliminate coupling with electromagnetic field outside of it .
mode spectrum of a perfect photonic crystal in the region of a band gap for perpendicular incidence ( @xmath52 ) is shown in fig .
[ fdef0 ] versus frequency @xmath53 and the number of elementary cell @xmath50 in which there is the @xmath54 layer , for which the calculation has been made . for simplicity , it has been assumed that the crystal is surrounded by a uniform medium with refractive index @xmath54 . as it can be seen in fig .
[ fdef0 ] , for elementary cells near the middle of the crystal mode spectrum is practically equal to @xmath51 in the gap with maximal values at both edges of the gap . outside
the gap mode spectrum fluctuates around the free space value .
it means , that the layers are resonators with resonant frequencies at the gap edges . for layers near the crystals edge
there is only a slight difference between mode spectrum of the layer and a uniform medium ( which is constant in the latter case ) , because they are resonators with a very low quality and do not impose a significant change of an amplitude of a plane wave . introducing a defect of width to a layer results in an appearance of a defect mode .
the size of the defect determines at which frequency does it happen see fig .
[ deffreq ] obviously , because a resonant frequency of a resonator depends on its width . however , the defect mode appears only in a certain number of layers near the defected one , because electromagnetic field with frequency from a band gap has an envelope which vanishes in consecutive layers it has a limited range of penetration , hence the defect has no noticeable influence on mode structure of layer which is sufficiently distant .
this behaviour is depicted in fig .
[ fdef1 ] ( values on @xmath10-axis in this figure are limited ) . and number of elementary cell @xmath50 . ] and ( b ) number of elementary cell @xmath50 for defect in @xmath55 elementary cell . ] figure [ fdef1max ] shows maximal values of mode spectrum for each frequency and in each layer .
it can be seen , that the defect introduces much higher values of mode spectrum than the edges of the band gap .
modification of mode spectrum depends on which layer is defected a similar graph for a defect in the middle of the photonic crystal is shown in fig .
[ fdef1maxcen ] . in this case
the defect mode frequency is the same , but values of mode spectrum associated with the defect are even higher . and ( b ) number of elementary cell @xmath50 for defect in @xmath56 elementary cell . ] and number of elementary cell @xmath50 . ] and ( b ) number of elementary cell @xmath50 . ]
if a photonic crystal has two defected layers , then there appear two regions in which influence of defect modes is significant .
overlap of these regions results in two defect modes , with two different frequencies , instead of one .
this case is shown in fig .
[ fdef2 ] and maximal values of mode spectrum are plotted in fig .
[ fdef2max ] .
adding more defect leads to further splitting of the band gap by a new defect band see figs .
[ fdef5 ] and [ fdef5max ] . and number of elementary cell @xmath50 . ] and ( b ) number of elementary cell @xmath50 . ]
in this paper we have analyzed the influence of defects in a 1d photonic crystal on the density of states and spontaneous emission , in both spacial and frequency domains .
we have used the effective resonator model and derived an analytical expression for mode spectrum . with the help of this relation
the presented numerical results revealing properties of 1d photonic crystal with defects have been obtained . in particular
, it has been shown that a single defect has local character and its influence on mode structure diminishes with the distance from the defect position .
project granted by the ministry of education and science in poland , project no 3 t11b 069 28 .
the authors are members of the sixth european union framework program for research and technological development ( fp6 ) network of excellence on micro - optics `` nemo '' . | the aim of this paper is to present the analysis of influence of defects in 1d photonic crystal ( pc ) on the density of states and simultaneously spontaneous emission , in both spatial and frequency domains . in our investigations
we use an analytic model of 1d pc with defects .
our analysis reveals how presence of a defect causes a defect mode to appear .
we show that a defect in 1d pc has local character , being negligible in regions of pc situated far from the defected elementary cell .
we also analyze the effect of multiple defects , which lead to photonic band gap splitting . |
phase space formulations of quantum mechanics are used to describe microprocesses strongly influenced by their external environment , by representing mixed quantum states in terms of phase space quasiprobability distributions , which are used to express expectation values or classical energy densities and fluxes as phase space averages @xcite . despite the increase in complexity of these formulations compared to the schrdinger representation ( nonuniqueness of phase space quasidensities , doubling of variables and nonlocality of evolution equations )
, they prove worth studying even outside the context of the theory of open quantum systems , for a series of reasons .
the phase space is the conceptually natural setting of quantum mechanics , more so in the setting of its explicit correspondence to classical mechanics , _ the semiclassical regime_. all inherent invariance properties of classical mechanics which become naturally manifest in phase space are reflections of approximate quantum mechanical symmetries , explicit in its phase space representations . also , the problem of caustics ( singularities void of physical content caused by projecting the dynamics to the configuration space ) is resolved since they are not encountered as persistent obstacles toward global solutions in phase space formulations as they do in semiclassical considerations in configuration space .
finally , a unified approach to the two main classes of semiclassical states , wkbm ( wentzel - kramers - brillouin - maslov ) states and coherent states , becomes plaussible in phase space ( e.g. , @xcite ) .
_ the wave packet representation _ is a phase space representation intimately related _ to deformation quantization _
@xcite , proposed in the works of torres vega and frederick @xcite , and also harriman @xcite , chruscinski and moldawski @xcite , de gosson @xcite and nazaikinskii @xcite , to mention a few ; in its core there is the correspondence of pure classical states , i.e. , single phase space points , to isotropic gaussian wave packets , @xmath0 localized at that phase space point on the heisenberg scale , @xmath1 , so that @xmath2 and @xmath3 . in this representation ,
_ the phase space wave function
_ , @xmath4 , is defined as the coefficient of the isotropic gaussian wave packet superposition of the wave function , at a given phase space point , @xmath5 this coherent state resolution , _ the wave packet transform _ , also known as
the fourier - bros - iagolnitzer transform @xcite , is closely related to _ the bargmann transform_. the dynamics of the phase space wave function satisfies the nonlocal _ phase space schrdinger equation_. while the apparent drawback of this representation is that it can not account for microscopic systems in strong interaction with their environment , it is a simpler representation than the wigner formalism , which is quadratic in the wave function . in the wave packet representation , the phase space schrdinger spectral problem
is well understood , e.g. , the works of luef and de gosson @xcite , who have derived the eigenvalue equation departing from moyal s spectral equation .
however , understanding the corresponding _ initial value problem _ is at the stage of infancy
. the deeper understanding of the initial value problem , its asymptotic investigation toward a direct theory of semiclassical dynamics in phase space is a challenging and invaluable contribution to phase space quantum mechanics , unifying approaches taken from the field of fourier integral operators @xcite to the theory of the maslov canonical operator @xcite , etc .
important contributions in this directions have been made by oshmyan et al and nazaikinksii et al @xcite .
a fundamentel semiclassical approximation of quantum dynamics is _ the gaussian approximation _ ; the idea of approximating quantum evolution with the dynamics of _ an individual _ gaussian wave packet translated along an orbit is traced back to the foundational work of schrdinger @xcite .
isotropic gaussian wave packets provide natural semiclassical approximations of quantum states for free motion , as they are localized in phase space on the heisenberg scale @xmath1 , ` occupying ' a planck cell centered at @xmath6 , and exhibiting oscillations at the de broglie wavelength @xmath7 ( for a general account on coherent states , see @xcite ) .
heller et al @xcite argued for the use of both single isotropic and anisotropic gaussian wave packets as approximations to the propagation of initially gaussian wave packets , based on _ the nearby orbit approximation _ , while in @xcite , huber , heller and littlejohn showed that isotropic gaussian wave packet dynamics can stand as a generalization of complex phase wkbm semiclassical propagation . in @xcite , hagedorn showed that the an initially gaussian state retains its gaussian form within a certain timescale under quantum dynamics .
these ideas have been implemented in the solution of the more general problem of establishing an asymptotic solution of the schrdinger equation or the wave equation along a given curve , modulated by a gaussian profile , known as _
gaussian beams_. gaussian beams find applications in a range of physical problems , in optics , accoustics , etc ( see , e.g. , @xcite ) . in the case of anisotropic gaussian wave packets , the ansatz @xmath8 for some real phase @xmath9 , the anisotropy quadratic form , which satisfies @xmath10 , is shown to obey the dynamics of the initially nearby orbit approximation .
these dynamics are equivalent to a matrix riccati initial value problem for the anisotropy matrix @xmath11 , common to all gaussian beam asymptotic solutions @xcite . in @xcite ,
littlejohn generalized the works of heller et al .
on the thawed , or anisotropic , gaussian dynamics for generic initially localized states , by constructing a nonlinear quantum propagator , as an explicit composition of weyl and metaplectic operators .
closely related to heller s and littlejohn s works , maslov and shvedov @xcite suggested the dynamics of anisotropic gaussian wave packets , giving an alternative representation of the dynamics of the anisotropy quadratic form , called by some authors _ the maslov - shvedov complex germ _ , as a one parameter flow in the siegel upper half space , related to the weil representation of the metaplectic group therein .
the gaussian approximation , however , breaks down as localization is lost in an irreversible spreading at a certain timescale , an effect suppressed only for quadratic potentials .
added to the above , the fact that for quadratic potentials the evolution of coherent _ superpositions _ of phase space eigenfunctions result in expressions reminiscent of the evolution of single gaussian wave packets , hints the method of approximating the evolution of a quantum state _ by a superposition of semiclassically propagated gaussian wave packets_. the starting point of this approximation is the resolution of the identity in quantum state space in the overcomplete basis of coherent states , in particular isotropic gaussian wave packets ; @xmath12 one makes an explicit ansatz for the propagation of a single wave packet under the schrdinger propagator @xcite , @xmath13 .
the totality of approximating for the single gaussian dynamics amount to _ the initial value representations _ of quantum dynamics .
the first work in this direction was that of herman and kluk @xcite , who argued on the validity of approximating semiclassical evolution by analyzing wave functions by a multitude of nonspreading isotropic gaussian wave packets , their shape held rigid , modulated by some overall amplitude and phase factor , begining from the van vleck - gutzwiller approximation of the propagator .
more recently , rousse and robert @xcite setting off from a ` head down ' direction , assumed a semiclassical time evolution for generic initial data of the schrdinger equation , in terms of a fourier integral operator identified with the herman - kluk propagator , in order to justify this approximation on the basis of rigorous estimates for the asymptotic solutions .
in this paper we turn attention to semiclassical asymptotic solutions of the initial value problem of the phase space schrdinger equation , rather than giving phase space representations of solutions of the schrdinger equation .
having defined the phase space quantum flow and phase space schrdinger propagator , we construct a semiclassical propagator , based on the anisotropic gaussian approximation , which is closely related to the nearby orbit approximation .
as the anisotropic gaussian approximation is made for the totality of orbits , all of which are taken into account in the wave packet resolution of the phase space wave function , the semiclassical propagator admits generic semiclassical initial data , not just localized ones .
remarking on the underlying metaplectic structure of the semiclassical phase space propagator , we construct asymptotic solutions of the phase space schrdinger for phase space wkbm initial data , as a half density semiclassically localized on the complex almost analytic extension of the lagrangian manifold of the solution of the corresponding hamilton - jacobi equation .
we consider hamiltonian systems and their quantization in configuration space @xmath14 , with cartesian coordinates @xmath15 . as a trivial vector bundle , the phase space @xmath16 , has the structure of a linear symplectic space with canonical coordinates @xmath17 and complex pseudocoordinates would be a symplectic pseudocoordinate transformation for the normalization @xmath18 , so that @xmath19 . ]
@xmath20 , where @xmath21 . for the action of the hamiltonian flow @xmath22
, we use the notation @xmath23 for the terminal point of the orbit @xmath24 eminating from @xmath6 . as the flow
is taken to be generated by an autonomous hamiltonian , generic apart from smoothness assumptions , including chaotic hamiltonians , we assume short time semiclassical propagation , @xmath25 , within _ the ehrenfest time _ ( e.g. , see @xcite ) .
_ the wave packet representation _ is an equivalent representation of quantum mechanics set in phase space , as the position representation is set in configuration space @xcite . by the stone - von neumann theorem @xcite ,
the two representations are related by means of a linear , unitary map , _ _ the wave packet transform _ _ into a phase space realization in @xmath26 , are equivalent modulo phase @xcite .
this can be verified by the uniqueness , up to phase , of the kernel of the wave packet transform @xcite . ]
@xcite , which maps configuration space wave functions @xmath27 to phase space wave functions @xmath4 , @xmath28 it is given explicitly by the integral transform @xcite @xmath29 the kernel of the transform is the isotropic gaussian wave packet with base point @xmath30 , complex conjugated , @xmath31 the wave packet transform is an isometry , in the sense that if @xmath32 is normalized to unity , then @xmath33 while @xmath34 . as a linear map between hilbert spaces ,
the wave packet transform is not ` onto ' ; its image is a subspace of @xmath26 , _ the fock - bargmann space _ ( see ( @xmath35 ) of the appendix ) ,
@xmath36 characterized by _ the fock - bargmann constraint _
@xcite @xmath37 or @xmath38 implying that _ not all square integrable phase space functions constitute phase space wave functions _
the fock - bargmann space @xmath39 possesses special analyticity properties ; it is essentially the space of gaussian weighted square integrable analytic functions ( see subsection @xmath40 of the appendix ) .
the wave packet transform microlocalizes quantum states in phase space ; the real part of the phase of the kernel contributes to semiclassical localization in the momenta , while the imaginary part to semiclassical localization in the positions . for @xmath41
, the inverse wave packet transform is defined as @xcite @xmath42 we consider the semiclassical schrdinger initial value problem corresponding to a mechanical system described by the hamiltonian function @xmath43 , @xmath44 the feynman indices stating the order of action of the symbol noncommuting operator arguments .
the dependence of the initial data on @xmath450,1]$ ] is assumed essentially singular at @xmath46 .
for the standard form , @xmath47 , the above problem takes the standard schrdinger form , for which operational ordering is unambiguous @xmath48 where @xmath49 is the positive laplacian in @xmath14 .
we consider semiclassical time evolution in the above problem for short times , which is quantified as @xmath50 , i.e. , within the ehrenfest time scale .
this is assumed in order to ensure the validity of the semiclassical approximation in the case the underlying hamiltonian flow is chaotic @xcite .
the phase space representation of the heisenberg quantized hamiltonian , @xmath51 , is ( see @xcite and @xmath52 of the appendix ) @xmath53 therefore , the wave packet transform of the initial value problem ( @xmath54 ) , the initial value problem for _ the phase space schrdinger equation _ reads @xmath55 for the phase space wave function @xmath56 . for for the standard form @xmath57 with @xmath58 the positive position laplacian in the phase space .
the wave packet transform retains the essential singularity of the equation itself as well as of the initial data , as will be shown in section ( [ solutions ] ) for wkbm states as prototype semiclassical states .
this renders the phase space image of the initial problem , in complete analogy , semiclassically singular .
a qualitative difference from the schrdinger equation in the case of the standard hamiltonian form arises from the potential term ; for nonpolynomial potentials , this term is a pseudodifferential operator , which introduces nonlocality as well as transport effects in phase space dynamics , both common features in phase space evolution equations , such as for the von neumann equation . in the case
the potential is real analytic , this pseudodifferential operator term is defined by the taylor series @xmath59 while in general , we have the representation @xmath60 the solution of the schrdinger initial value problem ( @xmath54 ) , with generic initial data @xmath61 , is given by the action of the schrdinger time evolution operator ( quantum flow ) on the initial data , @xmath62 @xmath63 being _ the schrdinger propagator _ , the schwartz kernel of @xmath64 . by beginning with the completeness relation of the isotropic gaussian wave packets , a resolution of the identity in @xmath65 @xcite , @xmath66 we acquire a phase space resolution of the schrdinger propagator @xmath67 by applying the wave packet transform ( @xmath68 ) on the representation formula ( @xmath69 )
, we derive _ the phase space quantum flow _
, @xmath70 , which yields the phase space wave function @xmath71 @xmath72 being _ the phase space schrdinger propagator _ , the schwartz kernel of @xmath70 . the phase space schrdinger propagator @xmath72 is expressed in terms of the schrdinger propagator , @xmath63 , by @xmath73 it can be easily checked that ( @xmath74 ) solves the problem ( @xmath75 ) , for generic initial data @xmath76 .
the above representation , @xmath77 serves as a starting point for _ gaussian approximations _ for phase space quantum dynamics ; as we shall see subsequently , these ammount to making a specific approximation for the dynamics of the single gaussian wave packet , @xmath78 ; the phase space propagator is , then , proportional to the wave packet transform of the evolved gaussian wave packet .
it follows that like time evolution in configuration space , time evolution in phase space is also generated by a unitary group @xmath79 constituting the phase space propagator @xmath70 a unitary quantum flow , in the sense that , on @xmath39 , @xmath80 so that phase space dynamics induced by @xmath70 retain the gaussian ` twisted ' analyticity of the initial data ( see relations @xmath81 and @xmath82 of the appendix ) .
as only those @xmath83 correspond to schrdinger wave functions on configuration space , only ` twisted ' analytic phase space functions correspond to _ pure quantum states _ , while all other choices of @xmath84 phase space functions correspond to _ mixed quantum states_. in terms of the phase space kernel , unitarity becomes manifest by the following @xmath85 and @xmath86 .
the schrdinger propagator @xmath63 is a weak solution of the schrdinger equation , @xmath87 taking the wave packet transform of the above , we have the analogous equation for the phase space propagator @xmath88 due to the gaussian integration in ( @xmath89 ) , the kernel @xmath72 has stronger smoothness properties than @xmath63 .
firstly , it has well defined values across the hyperplane @xmath90 , while for @xmath91 it does not converge weakly to a dirac distribution , but rather is a dirac mollifier on the heisenberg scale ( see ( @xmath92)-(@xmath93 ) of the appendix ) , @xmath94 which we define as @xmath95 . as an operator acting in @xmath26 ,
the initial phase space propagator is the kernel of the projector ( see ( @xmath96 ) of the appendix ) onto the subspace @xmath39 , @xmath97 meaning that if @xmath98 , then @xmath99 so that for @xmath100 the representation of the solution @xmath101 reconstructs the initial data , @xmath102 , @xmath103 by utilizing the completeness property .
this reproducing property of the initial propagator is a result of the underlying analytic structure of the fock - bargmann space @xmath39 , something which is elaborated on in the appendix ( subsection @xmath40 of the appendix ) .
in this section we deal with the propagation of gaussian wave packets in configuration space and their phase space image under the wave packet transform . in particular , we consider _ the anisotropic gaussian wave packet _ , following the independent works of heller , huber , littlejohn ( isotropic case ) @xcite and littlejohn @xcite , maslov , shvedov @xcite ( anisotropic case ) and more recently robert @xcite , nazaikinskii , schulze , sternin @xcite , and stoyanovsky @xcite .
an excellent account of the subject , masterfully touching upon its dynamical and algebraic aspects , is given by littlejohn @xcite .
as we shall see , the semiclassical dynamics of anisotropic gaussian wave packets in configuration space realizes a representation of the metaplectic group on the siegel upper half space , and is equivalent to the variational system of the hamiltonian flow in the frame of _ the nearby orbit method _ ( see , e.g. , @xcite ) .
an attempt for the semiclassical description of the dynamics in the position representation has come under the name of _ initial value representations _ , involving superpositions of initial semiclassical states in wave packets , @xmath104 , propagating them in a certain approximation along the hamiltonian orbits @xmath24 eminating from @xmath6 ( see introduction ) , and then superposing the evolved wave packets back together , with respect to _ all _ initial phase space points @xmath6 .
here we consider _ the anisotropic gaussian wave packet approximation _ , in order to construct a semiclassical propagator for the phase space schrdinger equation .
the leading term in the semiclassical propagation of a single normalized isotropic gaussian wave packet , amounts to the anisotropic gaussian approximation , namely @xmath105 retaining its gaussian form , yet acquiring an anisotropic phase .
in particular , in this approximation , the initial state @xmath106 is evolved semiclassically to the anisotropic gaussian wave packet moving along the ray @xmath107 , and semiclassically concentrated on that point on the heisenberg scale @xcite . in particular , we consider the following ansatz @xmath108 where @xmath109 is a real valued phase , and the matrix @xmath110 is symmetric and has positive definite imaginary part , @xmath111 and @xmath112 , to be determined by the dynamics . by substituting the ansatz ( @xmath113 ) in the schrdinger equation , and demanding that it be an asymptotic solution , in the sense that for a certain time scale and uniformly in @xmath114 , @xmath115 we arrive at a system of characteristic differential equations along the orbit @xmath24 @xcite , analogous to the wkbm system ; the zeroth order equation in @xmath116 is @xmath117 along with the initial condition @xmath118 , where @xmath119 is the phase space lagrangian of the flow .
these give the phase space action along the orbit @xmath24 , @xmath120 the following properties of the phase space action shall become of use in what follows , @xmath121 where appropriate matrix multiplication and contraction is implied . the first order equation in @xmath116 yields hamilton s equations for @xmath122 , which are automatically statisfied along the orbit @xmath24 .
the second order equation gives the amplitude @xmath123 and the anisotropy form @xmath11 .
the amplitude is @xmath124 offering an additional phase , or index to be accurate , @xmath125 the dynamics of the anisotropy form , sometimes called _ the maslov - shvedov complex germ _ , _ the matrix riccati equation _
@xcite , essentially controling the direction , the shape and the spreading of the propagated state , @xmath126 with the initial condition @xmath127 .
the time dependence of the anisotropy form is implicit in the flow @xmath128 , as the matrix riccati dynamics is taken along the orbit @xmath24 ; the anisotropy form @xmath11 depends on time only through @xmath122 , except for the class of quadratic hamiltonians which yield give solutions @xmath129 independent of the initial point @xmath6 .
we thus make the notational convention for the functional dependence @xmath130 , always for nonquadratic potentials .
the wave packet character of @xmath131 is guaranteed , as by beginning at @xmath127 , @xmath11 remains symmetric with positive definite imaginary part for all times .
we stress that _ the riccati equation guarantees unitarity of the semiclassical flow @xmath132 _ , in the sense that @xmath133 , for all times @xmath134 .
as by assumption @xmath135 , we must have @xmath136 and @xmath137 , the riccati equation must define a flow in _ the siegel upper half space _
@xmath138 which is a @xmath139-dimensional manifold generalizing the complex upper half plane @xmath140 @xcite .
the anisotropic gaussian wave packet is a semiclassical solution of the schrdinger equation for short times , @xmath141 , within the ehrenfest time scale , with initial data @xmath142 the short time scale is required so that there is no sufficient spreading , sustaining localization on the heisenberg scale is not lost . by semiclassical solution relation ( @xmath143 ) is meant , while there exists some @xmath144 such that @xcite @xmath145 and @xmath146 as noted in @xcite , in anisotropic gaussian wave packet dynamics the spatio - temporal variation in its localization is dictated by a certain matrix riccati dynamics , which in turn is related to the separation dynamics of initially nearby hamiltonian orbits .
these dynamics are essentially common to all gaussian narrow beam dynamics , i.e. , dynamics for asymptotic solutions concentrated with a gaussian profile about the characteristics @xcite .
the riccati equation of the anisotropy form is equivalent to the linear variational system governing the stability of the hamiltonian flow @xcite .
the hamiltonian flow is given by the system @xmath147 with initial data @xmath6 the stability of initially nearby orbits is given by the linearization of the dynamics _ about a reference orbit _ , rather than a fixed phase space point .
we note that the deviation coordinates from a given propagated point @xmath148 , say @xmath149 are canonical coordinate ( retaining the same notation for the original coordinates and the ` deviation ' coordinates ) , so that the nonautonomous linearized flow for the deviation , is itself hamiltonian @xmath150 where @xmath151 is the @xmath152 unit symplectic matrix .
the corresponding linear flow is given by a matrix symplectic transformation @xmath153 evaluated along the orbit @xmath24 .
the _ variational system _ , which relates the sensitivity of the system with respect to nearby initial points , is obtained by varying hamilton s equations by the initial points @xmath6 , @xmath154 where @xmath155 is the pseudocomplexified phase space pseudocoordinate ( see introduction ) .
following , e.g. @xcite , we write the above as @xmath156 where @xmath157 the variational system has a unique global solution , bounded away from zero for finite time @xmath158 @xcite .
the initial values are ( suppressing the phase space dependence ) , @xmath159 and @xmath160 .
with respect to the variational system , the anisotropy form @xmath11 is given by @xmath161 the initial condition @xmath162 determines the initial value @xmath127 , which is compatible with @xmath163 , so that there is an arbitrariness in the initial values of matrices @xmath164 and @xmath165 separately .
the dynamical representation of @xmath11 is , then , @xmath166 for hamiltonians of the standard form , @xmath47 , the variational system becomes @xmath167 and the matrix riccati equation @xmath168 for bounded motion ( compact energy shell ) and smooth potential , we have the following large time asymptotics , @xmath169 in matrix norm . finally , the amplitude becomes @xmath170 the dynamics of the anisotropy form as defined by the anisotropic gaussian wave packet approximation , is underlied by a rich algebraic and geometric structure .
the semiclassical evolution of @xmath131 is a representation of a one - parameter flow of @xmath11 in @xmath171 , _ the weil representation of the metaplectic group @xmath172 acting on _ @xmath171 , intimately related to the nearby orbit approximation , as described above @xcite .
as we described above , the nearby orbit method effectively approximates the hamitlonian flow by the linear flow generated by the quadratic hamiltonian @xmath173 fixed on a reference point @xmath148 .
under the quantized linearized flow induced by the above quadratic hamiltonian , one may construct an _ exact _ solution of the schrdinger equation with initial data @xmath174 if the linear flow , @xmath175 , is given by @xmath176 then the exact solution is @xcite , @xmath177^{-1}\kappa } \ , \ ] ] where @xmath178 , with @xmath179 .
this defines the action of the symplectic group on the set of gaussian wave packets , which simultaneously guarantees that the generalized mbius automorphism on @xmath171 @xmath180 is a group action of @xmath181 on @xmath171 .
thus , the class of anisotropic gaussian wave packets , @xmath182 which are obviously square integrable , @xmath183 transform into one another under the action of the symplectic group @xcite @xmath184^{-1}\kappa } \ .
\label{eq : spwp}\ ] ] although closely related to the symplectic group , the group action ( @xmath185 ) is not the representation of @xmath186 on @xmath65 gaussian wave packets , due to the double valuedness induced by the square root of the determinant factor @xcite . in order to achieve a single valued representation
, one must consider _ the metaplectic group _
, which is the double covering group of the symplectic group , taking into account exactly this ambiguity .
by representing a member of the symplectic group @xmath187 for appropriate @xmath188 and @xmath189 so that @xmath190 , we define the metaplectic group as @xmath191 the choice of the index @xmath192 being a choice of either branch of the square root of the determinant @xmath193 in the complex plane , for @xmath194 . @xmath195 the representation of @xmath172 on @xmath65 is called _ the weil representation_. its action on gaussian wave packets acquires the above simple form ; in general , given that the block @xmath196 , say , is nonsingular , the corresponding general metaplectic action on @xmath65 is given by @xcite @xmath197 assuming , without loss of generality that @xmath198 .
the above uncovers the following interconnection : semiclassical anisotropic gaussian wave packet dynamics is equivalent to a particular matrix riccati dynamics of the anisotropy quadratic form , which is identified with a one - parameter flow of @xmath172 in @xmath171 , which in turn , is equivalent to the stability dynamics of the hamiltonian flow . by establishing an approximate solution for the propagation of wave packets
, we can use the linearity of the schrdinger propagation and the wave packet transform to construct approximate solutions of the schrdinger equation , and in turn , of the phase space schrdinger equation , with any initial data .
the idea is trivial , intuitevely ; approximate the evolution of a general initial wave function by evolving the wave packets into which it is analyzed , and superpose them back together to obtain the asymptotic solution .
starting from ( @xmath199 ) , we note that a basic step toward the construction of a semiclassical wave packet propagator based on the anisotropic wave packet dynamics , @xmath200 is the transformation of wave packets into the phase space under @xmath201 . in particular , one must guarantee that the anisotropic wave packet dynamics retains , under @xmath201 , its gaussian form .
the wave packet transform of the anisotropic gaussian wave packet reads @xmath202 where @xmath155 , @xmath203 , and @xmath204 the overall phase is @xmath205 by ( @xmath206 ) .
the choice of the principal branch of the square root of the term @xmath207 is necessary in order to ensure agreement with the initial data @xmath208 , as @xmath209 , which in turn ensures agreement with generic initial data in for the action of the semiclassical quantum flow .
this choice of phase actually has a nontrivial algebraic content ; the double valuedness of the square root of the determinant reflects the double valuedness of the covering mapping @xmath210 , and thus choice of branch is actualy a choice of a specific metaplectic transformation . in order to verify the invariance of gaussian form under @xmath201 ,
as noted by folland @xcite , i.e. , that wave packets in configuration space are mapped into wave packets in phase space , one must bring the above in the form of a gaussian with a siegel quadratic form in @xmath6 , rather than in @xmath155 .
the auxiliary mapping @xmath211 , which will serve in defining the phase space anisotropy form as will be shown , is an action of the metaplectic group on @xmath171 , and its image is a submanifold of the open siegel upper half disk @xmath212 , @xmath213 and is actually identified with the open ball in the matrix norm topology , @xmath214 the closed ball is the image of the compactified siegel half space , @xmath215 , extending it by the real symmetric matrices ( the boundary ) and the point at infinity .
thus , we have the parametrization of the boundary by the matrix unitary group @xmath216 , @xmath217 in the light of this geometric restriction , the above is brought in a more transparent form by expressing the phase of @xmath218 in @xmath219 as a _ real _ quadratic form over phase space , @xmath220 the matrix of the quadratic form is @xmath221 the matrix @xmath222 as defined in @xmath223 .
this makes explicit the phase space gaussian wave packet character of @xmath224 , under the condition that the phase space anisotropy form guarantees square integrability , i.e. , @xmath225 . by noting , additionally , by the plancherel relation @xmath226 , that @xmath227 where @xmath228
, the gaussian form of @xmath218 becomes more apparent .
consider the mapping of the anisotropy form in phase space , @xmath229 as above . since @xmath230 is continuous , and since @xmath231 , all that is required in order to show that @xmath232 is to show that there are no points on the boundary @xmath233 , such that @xmath234 , where the manifold @xmath235 is parameterized by the equation @xmath236 this is indeed true , as the matrix @xmath237 where @xmath238 , is always nonzero .
thus , we see that the image of @xmath239 is mapped into a true subdomain of @xmath240 under @xmath230 , and so the dynamics of @xmath241 retains its siegel character .
thus , @xmath242 is a true phase space wave packet with anisotropy form @xmath241 .
this makes transparent the following geometric picture : _ the representation of the metaplectic group @xmath172 into @xmath65 is in 1 - 1 correspondence with the representation of the metaplectic group @xmath243 into fock - bargmann space , @xmath39 , by the wave packet transform_. in other words , the wave packet transformation of wave packets are true phase space wave packets themselves .
the fact that a gaussian wave packet retains its form in phase space is reinforced by the fact that its wigner transform , is in turn a real gaussian wave packet , bearing in mind that the modulus square of the phase space wave function is the convolution of the wigner function ( see section ( [ wignr ] ) of the appendix ) with a localized gaussian @xcite @xmath244 where @xmath245 defines a symmetric and positive definite quadratic form , related to @xmath241 .
as in the case of the anisotropic gaussian wave packet in configuration space , @xmath131 , the amplitude of its phase space image , @xmath218 , is controlled essentially by the quadratic term , which vanishes on the set @xmath246 , so that it becomes semiclassically concentrated there , exponentially small at a fixed distance away .
we now construct a semiclassical propagator for the phase space schrdinger equation , _ the semiclassical wave packet propagator _ , and a semiclassical flow , @xmath247 , which is based on the anisotropic gaussian wave packet approximation . by the inverse wave packet
transform ( @xmath248 we have @xmath249 and combining with ( @xmath250 , we get @xmath251 as @xmath252 is a semiclassical solution of the initial value problem ( @xmath253 with initial data @xmath254 , we have the approximation @xmath255 this , which is essentially the thawed gaussian approximation , is the central approximation we make to construct the semiclassical phase space propagator .
the last two equations yield the approximate solution @xmath256 the kernel @xmath257 is expressed as @xmath258 the kernel of the semiclassical flow @xmath259 .
we are interested in a semiclassical approximation for the phase space propagator . by substituting the approximate propagator @xmath257 into the representation formulae ( @xmath74 ) and ( @xmath89 )
, we obtain @xmath260 we term this _ the semiclassical wave packet approximation _ for the propagator of the phase space schrdinger equation .
the kernel of the semiclassical quantum flow @xmath261 is @xmath262 and is proportional to @xmath218 , @xmath263 so that from @xmath219 we finally deduce @xmath264 the semiclassical propagator @xmath265 is of the form of an approximate transition amplitude between a quantum state microlocalized at the phase space point @xmath6 and another moving with a time decay along the orbit @xmath266 , semiclassically localized on the manifold @xmath246 . by construction ,
the semiclassical propagator is a flow , in the group sense , over the set of wave packets ( in configuration space ) , @xmath267 for @xmath268 , which can be generalized to an arbitrary state by the overcompleteness of coherent states .
thus , we have that the approximate dynamics in phase space enduced by the operator @xmath261 , besides satisfying the fock - bargmann constraints , is a unitary quantum flow in @xmath39 , @xmath269 in terms of the semiclassical wave packet phase space propagator , we have @xmath270 and @xmath271 at @xmath100 , the semiclassical wave packet phase space propagator shares the reproducing property of the exact propagator , @xmath272 in @xcite , littlejohn constructed a semiclassical propagator based on the nearby orbit approximation , as an explicit action of weyl and metaplectic operators , generalizing the nearby orbit approximation for the dynamics of liouville densities , in the quantum mechanical framework .
this construction , however , does not fall in the category of initial value representations .
the littlejohn propagator utilizes a single reference orbit , with starting point the phase space point on which the initial data @xmath273 is assumed to be centered at and localized , without necessarily being gaussian .
although reminescent of single semiclassical gaussian wave packet propagation , it loses unitarity and linearity as the base point @xmath274 depends implicitly on @xmath273 .
the semiclassical littlejohn flow in the nearby orbit approximation reads @xmath275 where @xmath276 is the weyl operator ( see subsection ( [ twpt ] ) of the appendix ) , and @xmath277 defines the reference orbit with initial point @xmath278 .
the ambiguity of sign of the metaplectic operator is resolved by the requirement that @xmath279 , so that @xmath280 , while the symplectic transformation @xmath281 , is the solution of the initial value problem @xmath282 and the additional phase is @xmath283 .
the propagator fixes the initial state from @xmath274 to the origin , symplectically ` rotates ' it in phase space by the action of @xmath284 , and shifts it along the reference orbit modulating by adding the action phase , modulating with the extra action phase .
explicitly , we have @xmath285 although the method of constructing the semiclassical propagator @xmath265 in the previous subsection relies on the nearby orbit scheme as well , there is a number of differences to the littlejohn semiclassical propagator .
the littlejohn semiclassical propagator is simpler computationally , as it involves a single orbit eminating from the center of a microlocalized initial state .
subsequently , it does not grasp the quantum dynamics for nonmicrolocalized initial data ; in the anisotropic wave packet approximation elaborated on previously , one can consider initial gaussians for all possible initial phase space points , and thus all possible reference orbits in phase space .
this renders the use of semiclassical wave packet propagator valid even in the case of nonmicrolocalized initial data , such as wkbm states , concentrated on a certain lagrangian manifold .
the other differences have to do with the linearity and unitarity properties .
a different approach , on the lines of which we constructed the semiclassical wave packet propagator , is that of initial value representations for solutions of the schrdinger equation . by beginning with the completeness relation of the isotropic gaussian states in phase space @xmath286 , @xmath287 where @xmath63 is the schrdinger propagator ,
the following approximate phase space resolutions are usually made in this context , @xmath288 the above , are respectively _ the thawed gaussian approximation _ , _ the frozen gaussian approximation
_ , and _ the herman - kluk approximation _ @xcite .
we derive asymptotics for the schrdinger representation of the semiclassical wave packet propagator by means of the complex stationary phase method ( see subsection @xmath289 of the appendix ) . as was previously noted , @xmath290 bringing the above in the standard oscillating integral form , we have @xmath291 where the phase and amplitude read @xmath292 and @xmath293 the leading order contribution comes from the real critical manifold , @xmath294 which is given by the equations @xmath295 and @xmath296 as the matrix @xmath297 is positive definite , the later is possible only if @xmath298 and @xmath107 .
subsequently , the former stationary equations are , implying appropriate matrix muliplication , @xmath299 which by the properties of the phase space action @xmath300 are automatically satisfied .
thus , the real critical manifold becomes @xmath301 which corresponds to orbits , which , projected onto configuration space , yield rays beginning at @xmath302 and terminating at @xmath114 . as there are more than one such possibilities , the critical set comprises of a countable set of rays . on @xmath303
the phase itself reduces to the action , while the restriction of its hessian therein reads @xmath304 by virtue of the matrix poisson relation , @xmath305 the determinant of the hessian on the real critical manifold @xmath306 gives @xmath307 assuming short times , by the stationary phase lemma ( in the weak sense ) , @xmath308 where @xmath309 is the canonical projection down to configuration space , and @xmath310 is the set of orbits with conjugate points at @xmath302 and @xmath114 at times @xmath311 and @xmath158 respectively .
the phase @xmath312 is the corresponding maslov index of orbit @xmath313 , @xmath314 i.e. , the index of the monodromy matrix , the excess of its positive over negative eigenvalues , while @xmath315 is the action of the orbit @xmath313 .
the factor @xmath316 stands as a correction to the semiclassical van vleck - gutzwiller approximation of the schrdinger propagator . indeed , for @xmath91 , we have @xmath317 additionally , in order to ensure that @xmath257 serves as an approximation to the schrdinger propagator , we must have that the amplitude @xmath318 is nonsingular , even on the caustics .
the exponential factor @xmath319 is always bounded due to the global existence of solutions of the matrix riccati equation @xcite . as the flow @xmath320 constitutes a canonical transformation for all times , the determinant of @xmath321 is nonsingular , and thus @xmath322 remains bounded both away from zero and infinity for all times .
thus , we have shown , that for small times , the semiclassical wave packet propagator gives indeed the correct semiclassical asymptotics , @xmath323 the semiclassical van vleck - gutzwiller approximation of the propagator .
we now turn to the issue of central interest , the asymptotic solutions for the semiclassical initial value problem of the phase space schrdinger equation .
we construct such asymptotic solutions by semiclassically evolving appropriate phase space states under the semiclassical wave packet flow , @xmath261 .
as we consider asymptotic solutions in the semiclassical approximation , it is necessary to consider _ semiclassical initial data _ in phase space . to this end
we consider the phase space image , under @xmath201 , of the wkbm state @xmath324 the phase is real valued and smooth , @xmath325 , while the amplitude is real valued and rapidly decrasing , @xmath326 , satisfying , additionally a normalization condition @xmath327 the phase space wkbm state is , for any @xmath328 @xcite @xmath329 where @xmath330 and @xmath331 the tilde stands for _ the almost analytic extension _ , and @xmath332 stands for _ the almost analytic solution _ of the stationary equation @xmath333 , where @xmath334 is the complexification of @xmath114 , ( see subsection of the appendix ( [ almost ] ) ) , which reads , @xmath335 in a phase space neighborhood of the manifold @xmath336 , the lagrangian manifold of the initial phase @xmath337 , @xmath338 the imaginary part of the phase @xmath339 is non - negative , @xmath340 @xcite , becoming zero only for @xmath341 , the lagrangian manifold of the initial phase @xmath337 , so that @xmath342 . to leading order , the asymptotics of the wave packet transform of the initial data are @xmath343 assuming the principal branch of the square root . along
the lagrangian manifold , @xmath341 , we have @xmath344 considering the above class of phase space states as initial data , we construct the corresponding asymptotic solutions , @xmath345 , by the action of the semiclassical wave packet propagator @xmath346 which is asymptotic to the corresponding solution @xmath101 , in the sense that @xmath347 weakly , algebraically in @xmath348 . we examine the asymptotics of the integral representation of the asymptotic solution , according to the asymptotic series for @xmath349 in @xmath350 , which we write in normal form ( see subsection ( [ stplemma ] ) of the appendix ) , @xmath351 where @xmath352 and @xmath353 the critical manifold of the phase @xmath354 in complexified phase space , is @xmath355 where @xmath356 are complexifications of the phase space canonical coordinates ; the manifold @xmath357 is parameterized by the equations @xmath358 the general expression of the asymptotic solution is given by the complex stationary phase lemma ( see subsection ( [ stplemma ] ) of the appendix ) , @xmath359 as a phase space sumbmanifold , the nodal set of the imaginary part of the phase , @xmath360 ( see subsection ( [ almost ] ) of the appendix ) , @xmath361 is identified with the lagrangian manifold _ of the propagated phase _ , @xmath362 this is due to the fact that @xmath340 @xcite , and the positive definitiveness of the imaginary part of the phase space anisotropy form , @xmath225 .
thus , the real critical manifold becomes @xmath363 parameterized by the equations @xmath364 , or @xmath365 the asymptotic solution on the transported lagrangian manifold @xmath366 becomes @xmath367 one should note that in the leading order , the hessian of the phase of the critical manifold , i.e. its restriction on the real critical manifold , ammounts to the pullback of the flow , @xmath368 , so that @xmath369 since the time dependence of the anisotropy form is implicit in the flow , @xmath370 .
we are lead to the result @xmath371 then , according to the selection of the branch of the square root in subsection ( b.3 ) of the appendix , we have @xmath372 finally , we deduce that along the lagrangian manifold @xmath373 the asymptotic solution is @xmath374 we note that the transported phase constrained on the real critical manifold , @xmath375 where @xmath376 , is the solution of the corresponding _ hamilton - jacobi equation _
@xmath377 with initial data @xmath378 , so that @xmath379 . for reasons of convenience
we have assumed above that the lagrangian manifold @xmath366 is projectable , i.e. , it can be diffeomorphically projected onto the base space .
if this is not so , one can carry out the above arguments by introducing appropriate local coordinates , @xmath380 , so that canonically conjugate coordinate pairs are excluded , as was introduced by maslov in his construction of the canonical operator ( e.g. , @xcite ) .
this overcomes the problem of caustics due to folds of the lagrangian manifold @xmath366 . as the square modulus of the phase space wave function is a husimi density ( see subsection ( [ wignr ] ) of the appendix )
, we expect that in the classical limit it will give a singular measure , concentrated on the lagrangian manifold @xmath366 . for the asymptotic solution
, we have @xmath381 which by the generalized stationary phase lemma @xcite , becomes to leading order @xmath382 the induced riemannian metric tensor on the transported lagrangian manifold @xmath366 reads @xmath383 so that the riemannian measure on the manifold is @xmath384 , where @xmath385 having noted that @xmath386 , we deduce that @xmath387 in complete analogy to the wkbm asymptotic solution of the schrdinger equation . the difference , of course ,
is the existence of the jacobian factor related to the solution of the transport equation ( see , e.g. , @xcite ) ; lifted to the phase space , it is unity , the jacobian of a symplectomorphism , modlulo the geometry of the transported lagrangian manifold @xmath366 .
in order to put the semiclassical phase space propagator @xmath265 defined in @xmath388 , to the test , we compare the asymptotic solutions it produces against solutions of the phase space schrdinger equation for two simple idealized physical situations . in the first case we consider the propagation of a microparticle free of any external interactions , and in the other case scattering of a microparticle by a constant electric field , both in one dimension . in both cases we have subquadratic polynomial potentials , @xmath389 ,
so they share the common characteristic of linearity of the hamiltonian flow , and the phase space uniformity of the anisotropy form ; the matrix riccati equation is common for such systems , yielding the anisotropy form @xmath390 constant in the initial phase space points .
the phase space anisotropy form reads @xmath391 and @xmath392 while additionally , we have the phase @xmath393 we expect the semiclassical solution to be the exact solution , something characteristic for the class of quadratic potentials .
the quantum dynamics are given by the solution of the phase space initial value problem @xmath394 which stems from the configuration space problem , @xmath395 as initial data , for both cases , @xmath396 and @xmath397 , we set off by considering appropriate semiclassical initial data for the schrdinger equation , which allows for explicit calculations .
we take semiclassical initial data , a complex phase wkbm state @xmath398 so that the initial amplitude and phase are , respectively , @xmath399 and @xmath400 , generating the lagrangian manifold @xmath401 . the phase space image of the initial data @xmath273 under the wave packet transform is @xmath402 for @xmath155 .
the hamiltonian is @xmath403 which induces the flow @xmath404 , @xmath134 .
the phase space action along the flow is @xmath405 .
the solution is obtained by means of the free propagator @xcite , @xmath406 for the particular initial data , we obtain , @xmath407 which in phase space , becomes , by applying the wave packet transform , @xmath408 the semiclassical wave packet propagator is @xmath409\bigg ) \ .\end{aligned}\ ] ] the semiclassical solution by the semiclassical wave packet propagator , is given by @xmath410 for the particular choice of initial data above , the asymptotic solution constructed by means of the semiclassical wave packet propagator , @xmath261 , @xmath411 we unsurprisingly note that the asymptotic solution bore by the wave packet propagator , is actually identified with the exact solution .
the asymptotic solution , and thus the solution itself , is semiclassically concentrated on the lagrangian manifold as transported by hamilton - jacobi dynamics .
in particular , the hamilton - jacobi problem @xmath412 has the solution @xmath413 which induces the transported lagrangian manifold @xmath414 i.e. , initially the diagonal straight line @xmath401 , asymptotically tending to the horizontal straight line @xmath415 .
the hamiltonian is @xmath416 which induces the flow @xmath417 , @xmath134 .
the phase space action along the flow is @xmath418 .
the solution of the phase space problem is obtained by means of the airy propagator @xcite , @xmath419 for the particular initial data , we obtain , @xmath420 which in phase space , becomes , by applying the wave packet transform , @xmath421 ^ 2}{1-i+\hbar+2(1+i\hbar)t}\nonumber \\
+ \frac{i}{2}(q^2 - 2iqp)(1 + 2(1+i\hbar)t)-\frac{1}{3}t^3\bigg[1+\frac{1}{2}(1+i\hbar)t\bigg]\bigg ) \ .\end{aligned}\ ] ] the semiclassical wave packet propagator is @xmath422\bigg ) \ .\end{aligned}\ ] ] and so , the semiclassical solution by the semiclassical wave packet propagator , is given by @xmath410 for the particular choice of initial data above , we have , @xmath423(q - ip)+(1+i\hbar)q^2+\big[1 + 2(1+i\hbar)\big]tp^2 - 2i(1+i\hbar)qp\bigg ) \ .\end{aligned}\ ] ] as in the case of free motion , the asymptotic solution is identified with the exact solution .
the asymptotic solution , and thus the solution itself , is semiclassically concentrated on the lagrangian manifold as transported by hamilton - jacobi dynamics .
in particular , the hamilton - jacobi problem @xmath424 has the solution @xmath425 which induces the transported lagrangian manifold @xmath426 i.e. , initially the diagonal straight line @xmath401 , asymptotically tending to the horizontal straight line @xmath427 .
we have considered the singular initial value problem for the schrdinger equation in phase space and its semiclassical asymptotic solutions for wkbm phase space initial data .
we pass to the phase space representation by conjugating with the wave packet transform .
we introduce the phase space quantum flow , and the phase space propagator , and give a well posed initial value problem in the phase space .
the asymptotic solutions are constructed via a semiclassical propagator for the phase space schrdinger equation , the wave packet propagator , based on the anisotropic gaussian approximation .
the wave packet propagator defines a linear , unitary quantum flow on the fock - bargmann space , meaning that semiclassical quantum evolution preserves an initially pure state as such .
it is underlied by a rigid algebraic structure , characteristic of the anisotropic gaussian approximation and the nearby orbit approximation on which it is constructed .
it closely resembles the wave packet quantization of a canonical transformation @xcite .
the wave packet propagator is seen in a weak limit to approach the gutzwiller - van vleck propagator .
it is also tested against complex phase wkbm initial data , reproducing the exact solutions for the phase space initial value problem in the simple situations of free motion and scattering off a constant electric field .
the asymptotic solution produced by the semiclassical wave packet propagator yields a semiclassical husimi function on the lagrangian manifold @xmath428 , where @xmath337 is the initial wkbm phase ; alternatively , the lagrangian manifold generated by the solution of the hamilton - jacobi equation with initial data @xmath429 .
the relation to the maslov canonical operator can be seen by taking the inverse wave packet transform of the asymptotic solution @xmath430 , @xmath431 becoming asymptotically an integral over @xmath366 .
the authors would like to thank sergey dobrokhotov , frederic faure , maurice de gosson , robert littlejohn , vladimir nazaikinskii , vesselin petkov and roman schubert for fruitfull conversations and enlightening comments .
the wave packet transform defines a bounded operator @xmath432 whose image space , _ the fock - bargmann space _ , is a subspace of @xmath433 @xcite .
it satisfies the plancherel equality @xmath434 rendering it an isometric isomorphism , which guarantees the unitarity of the quantum flow in phase space and the conservation of the phase space probability @xcite . the wave function @xmath435 is represented in phase space by analyzing it in the overcomplete set of isotropic gaussian wave packets , @xmath436 .
isotropic gaussian wave packets constitute a nondenumerable , overcomplete basis in @xmath65 , in the sense that any @xmath437 can be written uniquely as a superposition of coherent states , which are , however , not orthogonal @xcite , @xmath438 rather , they are asymptotically orthogonal , @xmath439 for fixed @xmath440 , or , more specifically , @xmath441 while they fail to satisfy a completeness relation in @xmath65 , they satisfy a completeness relation in phase space @xcite , @xmath442 the main physical significance of coherent states is that they are the optimal wave functions in minimizing the product of the position - momentum uncertainties , naturally associating them with pure mechanical states , i.e. , points in phase space
. if @xmath443 is the multiplication operator by @xmath114 and @xmath444 , it is straightforward to incur the uncertainty relations while in the state @xmath445 , @xmath446 so that there is a saturation of the heisenberg inequality , @xmath447 @xcite .
the phase space wave function , @xmath448 , allows a statistical interpretation : _ is the probability amplitude of a microparticle being in the quantum coherent state _ @xmath449 .
it can not be interpreted in an analogous way to schrdinger wavefunctions , i.e. , the probability amplitude of a microparticle s position and momentum being @xmath6 , as the maximum resolution in phase space is the planck cell .
this reflects the fact that in phase space quantum theory the pure state is not related to a quantum state supported in a sense at that point , but a coherent state microlocalized on that point on the heisenberg scale . an equivalent way of defining the wave packet transform is by _ the weyl operator _
@xcite , @xmath450 where @xmath155 .
the weyl operator is unitary , and has the following property @xmath451 for @xmath203 , as well as the group structure , @xmath452 by the baker - cambell - hausdorff formula , it is simplified to @xmath453 .
the wey operator generates the set of isotropic gaussian wave packet by its action on the vacuum gaussian state , @xmath454 , @xmath455 its action on general gaussian wave packets is simply that of phase space translation , @xmath456 by means of the weyl operator , the wave packet transform can thus be expressed @xmath457 the following commutation relations of the wave packet transform with the conjugate canonical dynamical variables @xmath458 and @xmath459 hold @xcite : @xmath460 these relations define the wave packet image of the conjugate canonical pair , namely @xmath461 rearranging the above commutation relations , one readily incurs that any @xmath462 is constrained by the relation @xmath463 where @xmath464 , and @xmath465 is the annihilation operator , in terms of many body quantum mechanics . such a constraint on phase space wave functions comes as little surprise , as the doubling of the dimensionality of the base space introduces a great redundancy in the image space . in terms of the above constraint ,
_ the fock - bargmann constraint _
@xcite , the fock - bargmann space @xmath39 , as a subspace of @xmath26 can be written as @xmath466 making explicit its gaussian ` twisted ' analyticity property .
the fock - bargmann constraint defines an analyticity condition on the fock - bargmann space and the projector onto @xmath39 , which is elucidated by considering _ _ the complex fock - bargmann space _ _ , @xmath467 , which comprizes of analytic functions on @xmath468 , embedded in complex gaussian weighted @xmath84 space , @xmath469 , where @xmath470 is the gaussian , and @xmath471 the lebesgue measures on @xmath468 respectively .
here we have _ microlocalized _ symplectic complex pseudocoordinates , on the heisenberg scale , meaning @xmath472 reserving the same notation @xmath473 for convenience .
the above is made quite transparent by the form @xmath474 of the fock - bargmann constraint .
the isomorphism between the fock - bargmann spaces explicitly reads @xmath475 where @xcite @xmath476 with @xmath477 the linear symplectic phase space pseudocomplexification , @xmath478 the overall factor @xmath479 being a gaussian , @xmath480 the above isomorphism maps the fock - bargmann constraint on phase space wave functions to the cauchy - riemann analyticity conditions , @xmath481 , so that as a subspace of @xmath482 , the complex fock - bargmann space is @xmath483 the twisted analyticity properties of phase space wave functions , and their connection to the space of gaussian weighted phase space analytic functions , stems from the connection of the wave packet transform to _ the bargmann transform _ @xcite .
the wave packet transform is a ` semiclassical microlocalization ' of the bargmann transform .
explicitly , the analytic function determined by the isomorphism @xmath484 is the bargmann transform of @xmath435 localized on the heisenberg scale , @xmath485 the induced topological structures of the two spaces are given by the inner products @xmath486 and @xmath487 the adjoint operator of @xmath201 is @xcite @xmath488 where @xmath489 where @xmath490 is a sequence of bounded , compactly supported phase space functions with @xmath491 strongly in @xmath26 . the adjoint enables us to introduce the orthogonal projection @xmath492 @xcite , @xmath493 which plays a central role in the wave packet quantization formalism . restricted on @xmath39 , the adjoint @xmath494 is identified with the inverse @xmath495 @xcite , @xmath496 in particular , for @xmath41 @xmath497 there is a simple physical interpretation of the fock - bargmann decomposition of the phase space hilbert space , @xmath498 in phase space functions which are images of schrdinger wave functions , and those which are not .
the @xmath84 phase space functions outside @xmath39 account for _ mixed quantum states _ , which do not correspond to schrdinger wave functions , while the elements of @xmath39 account for _ pure quantum states_. the fock - bargmann projector @xmath499 induces equivalence classes of mixed quantum states @xmath500\subset l^2(\mathbb{r}^{2d})$ ] equivalent to the corresponding pure state @xmath501 , modulo scaling .
the local properties of @xmath499 are a smoothing on the heisenberg scale , smearing away oscillations on that scale @xmath502\big)}\psi ( \eta,\xi)\,dxd\eta d\xi \ .
\label{eq : berg1}\ ] ] for @xmath503 , there is a smooth kernel analogue of the dirac distribution customary in nonanalytic hilbert spaces . in particular , the kernel @xmath504 _ the bergmann reproducing kernel _ , has the reproducing property @xcite @xmath505 it is the integral kernel of the pseudocomlplexified fock - bargmann projector , @xmath506 , where @xmath507 in this subsection we make note of the relation between the heisenberg to the wave packet quantization .
we reserve the notations @xmath508 and @xmath509 for the two quantizations respectively . _ the heisenberg quantization of _
@xmath510 is defined as @xmath511 where @xmath512 is the multiplication operator by @xmath510 , and @xmath513 is the semiclassical fourier transform , @xmath514 . as a pseudodifferential operator , it reads @xmath515 while for analytic @xmath510
, we have the representation @xmath516 _ the wave packet quantization of _
@xmath510 is defined according to the rule @xcite @xmath517 in particular , we have @xmath518 a definition analogous to @xmath519 .
the difference can be understood in noting the rigidity in the order of integration in a fourier inegral operator representation of the later , @xmath520 according to definition @xmath521 .
the operator @xmath522 has the alternative form @xmath523 where @xmath524 is the projector onto the fock - bargmann space @xmath39 . as they ought to
, the above quantization rules share a common weak classical limit .
if @xmath525 , then we have the semiclassical relation @xcite @xmath526 differing , in general , even for polynomial symbols .
these operators can be expressed in the phase space representation , by means of conjugations with the wave packet transform .
one should , however , take care not to confuse _ the wave packet quantization of some physical quantity _
@xmath510 , which is an operator on an appropriate subspace of @xmath65 , with _ the phase space representation _ of some operator over @xmath65 , which is an operator over a subspace of @xmath26 .
the former characterization is concerned with a specific quantization scheme , while the later with a specific representation of operators in a different quantum state space . for any operator @xmath527
, we define its phase space representation as @xmath528 for heisenberg operators
@xmath529 , we have @xmath530 where we have used the relations @xmath531 .
we use the notation @xmath532 analogously , the phase space representation of the wave packet quantization of @xmath510 reads @xmath533 we note that the heisenberg and wave packet quantizations of the cannonically conjugate pair @xmath534 and @xmath535 are identical , @xmath536 which leads to the phase space representation @xmath537 for a detailed account on the semiclassical wave packet quantization of symplectomorphisms , the thrird essential ingredient of the physical theory next to mixed states and physical quantities , see nazaikinskii et al @xcite .
the wave packet quantization is of no use in the description of systems in strong interaction with their external environment . in this case ,
a wave fucntion description is not possible , and the quantum system is necessarily described by a statistical ensemble of wave functions .
the quantum state of motion is in this case a _ density operator _ , a trace class rank one projection operator of the form @xcite , @xmath538 over some countable wave function ensemble , @xmath539 .
@xmath540 is the orthogonal projector onto the ray @xmath541 , and @xmath542 are probabilities constrained by the normalization @xmath543 , reflecting the potential possibilities of the system being in a certain state @xmath435 .
the notation used @xmath544 does not imply that the density operator is a quantization of some classical phase space density . in the case of an isolated quantum system ,
the above reduces to @xmath545 the phase space representation of the density operator is @xmath546 the wigner function is @xmath547 the weyl symbol of the density operator @xcite . unlike the wave packet transform ,
the wigner transform defines a _ bilinear _ integral transform of the wave function .
it is also real valued , unlike the complex valued phase space schrdinger equation .
the dynamics of the density operator is defined by the heisenberg equation @xmath548=0\ .\ ] ] by taking the weyl symbol of the left hand side , we have _ the von neumann equation _ , or _ quantum liouville equation_. if the underlying classical system is generated by the hamiltonian @xmath43 , then the time evolution law for the wigner function reads @xcite @xmath549 here we have introduced the noncommutative , nonassociative moyal product @xcite , @xmath550 the moyal product induces the moyal bracket , @xmath551 which is a noncommutative , nonassociative deformation of the poisson bracket for positive @xmath348 . in terms of the moyal bracket
, the von neumann equation can be written in a form which makes its characterization as the quantum liouville equation more apparent , @xmath552 there is a direct connection between the wigner formalism and wave packet quantization . the density function @xmath553 which defines a phase space probability measure
can not be equated to the wigner function , as it assumes negative values as well , but rather to the corresponding _ husimi density _
@xcite , @xmath554 the convolution of the wigner function with a gaussian , @xmath555 , over the heisenberg scale , smearing away oscillations on that scale which sweep both positive and negative signs , rendering it a true phase space densitiy .
the most striking advantage of the wigner quantization , is that , in contrast to the wave packet quantization , it allows one to consider mixed quantum states . in other words ,
any phase space wave function in @xmath26 , along with mild smoothness conditions , are admissible initial data for the von neumann equation , whereas the phase space schrdinger equation admits only initial data in @xmath39 , corresponding to pure quantum states , if one is to consider it not merely as a mathematical pseudo - differential equation stripped of all physical content .
in the present we give an outline of the fundamental result on the asymptotic behavior of highly oscillating integrals with complex phase function , of the form resulting from the construction of wkbm wave functions .
in particular , a version of the method of stationary phase for complex phase functions . for arbitrary @xmath328 , we are led to the concept of _ the almost analytic extension _
, @xmath565 such that @xmath566 , @xmath567 , satisfying the cauchy - riemann conditions to all order ( @xmath568 here is not to be confused with the pseudocoordinate introduced in appendix a ) , @xmath569 for @xmath570 nearly real .
if @xmath510 is real analytic , @xmath571 coincides with its unique analytic continuation in @xmath572 .
following malther @xcite , a representation of the almost analytic extension of @xmath510 is @xmath573 where @xmath574)$ ] , supported inside the unit ball @xmath575 , becoming unity at a small neighborhood of the origin , @xmath576 , and @xmath577 if the fourier transform of @xmath510 , ( using the same notation as in the appendix for the semiclassical fourier transform ) @xmath578 in this subsection we follow the work of sjstrand and melin @xcite .
firstly , for a smooth function @xmath580 , we define the stratified manifolds @xmath581 to be _ the critical set _ and _ nodal set _ of @xmath510 , respectively .
we now consider the asymptotic behavior of oscillating integrals , which are of the form @xmath582 for small positive @xmath348 and @xmath583 .
the following conditions are assumed : + 1 ) the phase function is @xmath584 , where @xmath585 and @xmath586 are neighborhoods of the origins of @xmath587 and @xmath588 respectively , with nonnegative imaginary part , @xmath589 , equality being saturated at the origin @xmath590 ; the origin is also a nondegenerate stationary point , @xmath591 with @xmath592 , which leads to @xmath593 in @xmath594 .
the critical manifold of @xmath595 , @xmath596 , is thus nondegenerate .
additionally , the nodal set @xmath597 is identified with the critical set @xmath598 .
+ 2 ) the amplitude is @xmath599 , where @xmath600 , supported in a neighborhood of the origin within @xmath601 , admitting a regular semiclassical asymptotic expansion in the interior of @xmath602 , @xmath603 as @xmath510 is smooth , it admits an almost analytic extension @xmath571 into the complex domain @xmath604 , where @xmath605 containing the real subspace @xmath587 , satisfying the cauchy - riemann conditions to all orders , @xmath606 for @xmath570 nearly real .
let there be a real critical point , i.e. , @xmath607 such that @xmath608 . by the implicit function theorem
, the critical manifold @xmath609 continues locally into the complex domain @xmath572 , parameterized by the almost analytic coordinate chart @xmath610 always in a small neighborhood of @xmath611 , with @xmath612 , say the ball @xmath613 for some @xmath614 .
the above is termed an _ almost solution _ of @xmath615 .
the positivity of the imaginary part continued into the complex critical manifold is guaranteed by sjstrand @xcite , @xmath616 for @xmath617 , and some @xmath618 . by notation
, it should be clear that @xmath619 does not stand for @xmath620 . additionally , by the grding inequality
, we have @xmath621 for some different @xmath618 . by sjstrand ( @xcite , p. 145 ) ,
@xmath622 is equivalent to @xmath623 .
thus , the later inequailty implies that points belonging to the nodal set of the imaginary part of @xmath571 , @xmath624 , are necessarily critical points ; thus , we have @xmath625 however , by the bound for the imaginary part of @xmath571 , we have that any point @xmath617 with @xmath626 , can not belong to the nodal set @xmath627 , and thus , by the connectedness of @xmath628 , all zeroes of @xmath629 are necessarily real critical points , @xmath630 one must show that all real stationary points are necessarily zeroes of @xmath629 , i.e. , that the real restriction @xmath631 is identified with @xmath632 as manifolds in @xmath633 .
this requirement is met by the class of wkbm states as defined in the previous sections . + * theorem * ( sjstrand and melin @xcite ) _ let @xmath318 and @xmath510 be amplitude and phase functions as defined above .
then , there is a neighborhood @xmath585 and @xmath586 of the origins of @xmath587 and @xmath588 respectively , and differential operators @xmath634 of order less than @xmath635 , with smooth coefficients in @xmath636 , such that _ @xmath637 _ in @xmath6380,1])$ ] . in the above
, @xmath639 is the equation defining the almost analytic manifold @xmath609 in a complex neighborhood of @xmath601 , @xmath571 and @xmath640 are the almost analytic extensions of @xmath510 and @xmath641 in the same complex neighborhood .
the square root in the left hand side is the principal branch . _ in the above context , we call a point @xmath642 a _ real stationary point _ if there is a @xmath643 for which the following equations hold @xmath644 which , of course , implies that @xmath642 is itself real . we are lead to define _ the real critical manifold _ of @xmath510 as @xmath645
it is clear that for oscillating integrals of the above form , real stationary points -if any- give a non exponentially decaying contribution in @xmath348 , while genuinely complex ones will give exponentially small contributions in @xmath348 .
the solution of @xmath615 , which determines the stationary manifold , is @xmath646 . for @xmath647
close enough to the real stationary manifold , @xmath648 , we have @xmath649 , where @xmath650 parameterizes the real manifold at the point projected down by @xmath651 , and @xmath652 ( which is not purely imaginary ) is linear in the distance @xmath653 , for @xmath647 close enough to the manifold .
the real critical manifold approximation amounts to @xmath656 where @xmath650 is a parametrization of the real critical manifold @xmath648 .
the above approximation gives the same leading behavior for @xmath657 as the above almost analytic extension expansion .
there is no need to employ the almost analytic extension of @xmath510 , as @xmath658 for real @xmath114 .
the branch of the root @xmath659 is chosen so that @xmath660 , where @xmath661 are the eigenvalues of the hessian @xmath662 , with @xmath663 .
the quadratic term contributes @xmath664 , as the microscopic vicinity of @xmath648 is such that @xmath665 .
it is clear that in the case there are no real critical points @xmath666 , the imaginary part of the phase is nonzero , and thus there is an overall exponentially decaying behavior . explicitly , if @xmath667 , @xmath668 the integral is exponentially decaying everywhere in @xmath588 . + | we consider the singular semiclassical initial value problem for _ the phase space schrdinger equation_. we approximate semiclassical quantum evolution in phase space by analyzing initial states as superpositions of gaussian wave packets and applying individually semiclassical anisotropic gaussian wave packet dynamics , which is based on the _ the nearby orbit approximation _ ; we accordingly construct a semiclassical approximation of the phase space propagator , _ the semiclassical wave packet propagator_. by the semiclassical propagator we construct asymptotic solutions of the phase space schrdinger equation , noting the connection of this construction to _ the initial value repsresentations_. |
the signature of the @xmath12-process contribution to the solar abundances implies two parts , a @xmath13 component , which is responsible for the mass region from y to bi , and a @xmath14 component , which dominates in the region between fe and sr .
the main component can be assigned to low mass stars with @xmath15 , whereas the weak component is related to massive stars with @xmath16 ( @xmath17 stands for the mass of the sun ) . because of the much shorter evolution times of massive stars , the @xmath12-process enrichment of the universe starts with the lighter @xmath12 elements .
accordingly , the weak @xmath12 process is important for the early stellar populations and for galactic chemical evolution in general @xcite .
the @xmath12 process in massive stars occurs in two steps during different evolutionary phases .
neutrons are mainly produced by the @xmath18ne(@xmath19)@xmath20 mg reaction in both cases , but at rather different temperatures and neutron densities . during core he burning ,
neutrons are produced near core he exhaustion at temperatures of @xmath21=(2.5 - 3.5)@xmath22 k for about 10@xmath23 years with neutron densities @xmath24 @xmath25 , whereas the higher temperatures of @xmath21=(1.0 - 1.4)@xmath26 k during the subsequent carbon shell burning phase give rise to peak neutron densities of about @xmath27 @xmath25 @xcite . in the scenarios for the main and weak @xmath12 process ,
the stellar ( @xmath28 ) cross sections of the involved isotopes constitute the essential nuclear physics input , but with an important difference : the high neutron exposure during the main component leads to equilibrium in the reaction flow , expressed by the so - called local approximation @xmath29 , which holds true for isotopes between magic neutron numbers .
this relation implies that the emerging @xmath12 abundances are inversely proportional to the stellar cross sections and that the uncertainty of a particular cross section affects only the abundance of that specific isotope .
in contrast , the neutron exposure in massive stars is too small to achieve flow equilibrium , and this means that cross section uncertainties are not only influencing the abundance of that particular isotope but have a potentially strong propagating effect on the abundances of the subsequent isotopes involved in the @xmath12-process chain .
this propagation effect was first discussed for the @xmath30ni(@xmath31)@xmath32ni reaction @xcite and later investigated for the complete reaction chain of the weak @xmath12 process @xcite .
computations with different choices for the maxwellian averaged cross sections ( macs ) showed important bottle - neck effects and the significant impact of crucial capture rates near the fe seed on the abundances of the subsequent isotopes in the reaction chain .
limitations in the quality of stellar cross sections can , therefore , have serious consequences for the contributions of the weak @xmath12 process to galactic chemical evolution .
stellar neutron capture cross sections in the mass region of the weak @xmath12 process are often not available with the required accuracy . in case of the zn isotopes ,
experimental data suffer from large uncertainties or are even missing as for @xmath2zn @xcite .
therefore , a series of activation measurements has been performed at the karlsruhe 3.7 mv van de graaff on natural zn to improve the accessible cross sections for @xmath33zn and for the partial cross section of @xmath1zn .
because these isotopes have relatively small cross sections of less than about 50 mb , they are expected to give rise to large propagation effects in the final abundance distribution . apart from its propagation effect in the weak @xmath12 process , zn is of general interest for galactic evolution . though it is commonly accepted that zn is produced by a variety of scenarios the respective nucleosynthesis mechanisms and their relative contributions
are poorly understood .
spectroscopic observations over a range of stellar metallicities ( e.g. refs .
@xcite ) found indications of an overabundance of zn compared to fe in low - z stars , but an explanation for this excess appears premature in view of the present uncertainties concerning the origin of zn . in massive stars , the @xmath12 process component of the zn isotopes
is obscured by the fact that measured neutron capture cross sections are scarce , uncertain or even missing .
experimental technique and measurements are described in sec .
[ exp ] , secs .
[ ana ] and [ resu ] deal with data analysis and results .
maxwellian average cross sections ( macs ) and the related astrophysical aspects are discussed in sec .
[ astro ] .
the activation method represents a well established and accurate approach to determine macss at @xmath3 kev by producing a quasi - stellar neutron spectrum via the @xmath34li(@xmath35)@xmath34be reaction @xcite .
this method has been extensively used , mostly for measurements related to the main @xmath12-process component ( for examples see refs .
@xcite ) . in the present experiment ,
the proton beam with an energy of @xmath36=1912 kev , 30 kev above the reaction threshold , was delivered by the karlsruhe 3.7 mv van de graaff accelerator with typical intensities of 100 @xmath37a .
the neutron production targets consisted of 30 @xmath37 m thick metallic li layers evaporated onto water cooled copper backings . in this way
, neutrons are kinematically collimated into a forward cone of 120@xmath38 opening angle .
neutron moderation is avoided since cooling is achieved by lateral heat conduction to the water flow outside of this cone . throughout the irradiations
the neutron flux was continuously monitored and recorded in time steps of typically 60 s by means of a @xmath39li - glass detector at 1 m distance from the target .
this information is important to account for fluctuations of the neutron yield in evaluating the fraction of the reaction products that decay already during the irradiations . for a sketch of the experimental setup and more details
see ref .
@xcite for example . in total ,
four activations were performed with three metallic zn samples 8 and 12 mm in diameter and 0.5 and 1 mm in thickness to minimize uncertainties stemming from sample geometry ( table [ tab : samples ] ) .
the samples were sandwiched between 0.03 mm thick gold foils of the same diameter , which served as neutron flux monitors using the well known au reference cross section @xcite . during the activations
the samples were placed completely inside the neutron cone in contact with the neutron target at the position of highest flux .
the relative isotope abundance ratios adopted from ref .
@xcite are 48.268(214 ) for @xmath0zn , 19.024(82 ) for @xmath1zn and 0.631(6 ) for @xmath2zn .
.zn samples and irradiation parameters . [ cols="^,^,^,^,^,^ " , ] the experimental uncertainties are summarized in table [ tab : uncertain ] , where the investigated reactions are indicated by the respective product nuclei .
significant contributions to the overall uncertainty originate from the gold reference cross section , the efficiency of the hpge detectors , and the time integrated neutron flux .
the @xmath2zn cross sections are also affected by uncertain @xmath40-decay intensities .
any improvement of these data would , therefore , be important . because statistical uncertainties were found to be practically negligible after averaging the results from repeated activations , the final uncertainties are determined by systematic effects . only in case of the partial cross section to the ground state of @xmath8zn ,
the counting statistics was not sufficient for achieving an adequate accuracy of the final result .
lccccc source of uncertainty & + & au & @xmath5zn & @xmath6zn@xmath41 & @xmath8zn@xmath42 & @xmath8zn@xmath41 + gold cross section & 1.5 & - & - & - & - + number of nuclei & 0.1 & 0.5 & 0.4 & 0.9 & 0.9 + time factors , @xmath43 , @xmath44 & & 0.1 & 0.2 & 4.0 & 1.3 + self - absorption , @xmath45 & + detector efficiency , @xmath46 & + @xmath40-ray intensity , @xmath47 & 0.06 & 0.2 & 0.2 & 8.2 & 3.8 + time integrated flux , @xmath48 & - & 1.5 & 1.5 & 2.5 & 2.4 + counting statistics & & 0.7 & 0.9 & 13 & 0.8 + & & & & & + total uncertainty & - & 2.8 & 2.8 & 16 & 5.1 +
table [ tab : results ] shows a summary of the results obtained in all four activations together with the corresponding uncertainties . in spite of the variation of the experimental parameters ( see table [ tab : samples ] ) , the results are all consistent within the estimated uncertainties , thus confirming the procedures applied in data analysis .
these variations included different sample sizes and masses to verify the corrections for finite size and self shielding effects as well as different irradiation times to control uncertainties due to the half - life of the respective product nucleus .
@ccccc activation & @xmath40-ray energy & cross section & mean value + & ( kev ) & ( mb ) & ( mb ) + + i & 1115.5 & [email protected] & + ii & 1115.5 & [email protected] & + iii & 1115.5 & [email protected] & + iv & 1115.5 & [email protected] & [email protected] + & & & + + i & 438.64 & [email protected] & + ii & 438.64 & [email protected] & + iii & 438.64 & [email protected] & + iv & 438.64 & [email protected] & [email protected] + & & & + + i & 910.27 & [email protected] & [email protected]@xmath50 + & & & + +
i & 487.38 & [email protected] & + ii & 487.38 & [email protected] & + iii & 487.38 & [email protected] & + iv & 487.38 & [email protected] & [email protected] + + @xmath51cross section averaged over quasi - stellar spectrum .
+ @xmath50only activation i could be considered , all other activations lasted too long .
previous ( @xmath28 ) cross section data for the investigated isotopes @xmath52zn are rather limited @xcite . for @xmath0zn as well as for @xmath1zn
, there is only a single measurement based on the time - of - flight ( tof ) method . in both cases , uncertainties of 10%
have been reported @xcite .
the partial cross section to the isomer in @xmath6zn has been studied once @xcite and was quoted with an uncertainty of 25% , whereas there are no experimental results for @xmath2zn at all . as illustrated in table
[ tab : results ] , these uncertainties could be substantially improved .
the numerical cross section data are compared for the macs values in the following section .
the values listed in table [ tab : results ] represent average cross sections for the experimental neutron distribution used in the irradiations .
though the experimental spectrum corresponds in very good approximation to a thermal spectrum for @xmath53 kev , the cut - off at 106kev requires a small correction , in particular if the investigated cross section exhibits a different energy dependence than the gold reference cross section . for the calculation of the final macss @xmath54 the correction
was obtained by normalizing the differential ( @xmath55 ) cross sections , @xmath56(@xmath57 ) , e.g. from theoretical calculations with the hauser - feshbach ( hf ) statistical model @xcite or from evaluated data libraries ( http://www-nds.iaea.org/ ) , with the new experimental values .
apart from the normalization factor @xmath58 from the definition of the macs in eq .
[ macs ] , these corrections are between 3% and 5% . the normalized energy - dependent cross sections can also be used for extrapolation to other temperatures as shown in table [ tab : macs ] .
the data from the compilation of refs .
@xcite are listed before and after normalization to the present results for @xmath53 kev @lccccccccccc & + @xmath59 ( kev ) & 5 & 10 & 15 & 20 & 25 & 30 & 40 & 50 & 60 & 80 & 100 + & & & & & & & & & & & + + ref .
@xcite & 139 & 108 & 88 & 75 & 66 & 59@xmath495 & 52 & 47 & 44 & 40 & 38 + ref .
@xcite@xmath60 & 123 & 95.9 & 78.2 & 66.6 & [email protected] & [email protected] & 46.2 & 41.7 & 39.1 & 35.5 & 33.7 + & & & & & & & & & & & + + ref .
@xcite & 331 & 238 & 197 & 174 & 153 & 139@xmath496 & 121 & 113 & 102 & 87 & 79 + & & & & & & & & & & & + + ref .
@xcite & 57 & 38 & 30 & 26.3 & 23.5 & 21.5@xmath492 & 18.7 & 16.9 & 15.5 & 13.7 & 12.5 + ref .
@xcite@xmath60 & 28.4 & 18.9 & 14.9 & 13.1 & [email protected] & [email protected] & 9.3 & 8.4 & 7.7 & 6.8 & 6.2 + @xmath51 for comparison with measured cross section values in table [ tab : results ] multiply with @xmath61 + @xmath60 normalized to the measured value .
macs values are commonly compared at the standard thermal energy of @xmath62 kev ( table [ tab : macs ] ) .
the value for @xmath0zn is about 12% lower than the macs based on previous tof data @xcite , but the uncertainty could be reduced by a factor of 2.5 . in case of @xmath2zn , the sum of the partial cross sections is considerably smaller than the previously compiled value that had been obtained by an hf calculation and an additional empirical correction factor @xcite .
the factor of two difference found for the @xmath2zn cross section , which had been inferred from purely theoretical hf predictions @xcite , is reflecting the uncertainty of the hf approach in this mass region .
[ fig : network ] shows the @xmath12-process path in the region around zinc .
there are two potential branchings in the reaction flow affecting the production of the important @xmath12-only nucleus @xmath2ge .
the first branching bypassing @xmath2ge could occur between @xmath63-decay and neutron capture at @xmath6zn .
however this branching is not open , since the @xmath63-halflives of ground and isomeric state are too short ( 1 h and 14 h ) .
therefore , @xmath6zn will always undergo a @xmath63-decay before capturing a neutron .
the second potential branching could occur between @xmath63- and ec - decays at @xmath2ga .
however , the ec - decays occurs only in 0.4% of all decays , hence the branching can be neglected .
this means , the s - process path is not even partly bypassing the s - only nucleus @xmath2ge , hence @xmath2ge is a well suited normalization point between s - abundance and solar abundance . on the other hand ,
the @xmath12-process path is completely bypassing @xmath2zn , which makes it an @xmath64-only nucleus .
-process reaction network in the region around zn . ]
the @xmath12 process in the zinc region was investigated using the nucleosynthesis code netz @xcite .
the information on the respective stellar scenarios , i.e. the effective time dependent profiles for temperature , mass density , and neutron density in the considered burning zones were adopted from @xcite .
the cross sections used are the latest recommended values from the kadonis database @xcite , which followed the compilation of @xcite . the most important change in the abundance distribution of the weak @xmath12 process results from the new , reduced neutron capture cross section of @xmath0zn . the comparison of the overabundance factors in fig .
[ fig : sensitivity ] shows that the abundance of @xmath0zn is increased by 11% , while most of the heavier elements up until mass 90 are slightly less produced because of the aforementioned propagation effect that is typical for the conditions of the weak @xmath12 process .
this effect reflects the fact that the local equilibrium of the s - process is not reached during the conditions of the weak @xmath12-process component .
zn for the overproduction in the weak @xmath12 process .
@xmath2ge is used as a normalization point of the @xmath12-process abundances .
because @xmath2zn is bypassed during the weak @xmath12-process , its strongly reduced cross section has no significant impact .
this holds true also for the neutron rich isotopes of germanium and selenium ( @xmath65ge , @xmath66se ) . ]
the neutron capture cross sections of @xmath0zn , @xmath1zn , and @xmath2zn have been measured with the activation technique in a quasi - stellar neutron spectrum corresponding to a thermal energy of @xmath3 kev .
samples of natural isotopic composition have been used . by a series of repeated irradiations with different experimental conditions , the @xmath0zn(@xmath4)@xmath5zn cross section was determined to [email protected] mb and the partial cross section @xmath1zn(@xmath4)@xmath6zn@xmath7 feeding the isomeric state in @xmath6zn to [email protected] mb for the experimental spectrum .
values of [email protected] mb and [email protected] mb were found for the partial cross sections of @xmath2zn(@xmath4)@xmath8zn@xmath7 and @xmath2zn(@xmath4)@xmath8zn@xmath9 , which had not been measured so far .
the half - life of @xmath8zn@xmath41 could be determined to [email protected] h , with a significantly improved accuracy compared to the previous recommendation .
the weighted average of the combined data yields a value of [email protected] h. compared to previous measurements on @xmath10zn , the uncertainties could also be significantly improved , while the @xmath2zn cross section was found two times smaller than existing model calculations . from these results
maxwellian average cross sections were determined between 5 and 100 kev .
the consequences of these data have been studied by network calculations based on the temperature and neutron density profiles for convective core he burning and convective shell c burning in massive stars .
these simulations of the weak @xmath12 process reveiled that only the change of the @xmath0zn(@xmath4)@xmath5zn cross section results in significant abundance changes compared to the previous set of macs values .
the other cross sections have much smaller impact because the macs of @xmath1zn did not change and @xmath2zn lies outside the main path of the weak @xmath12-process component .
we would like to thank e .-
knaetsch , d. roller , and w. seith for their support at the karlsruhe van de graaff accelerator .
this work was supported by the hgf young investigators project vh - ng-327 and the eurogenesis project masche . | the neutron capture cross sections of @xmath0zn , @xmath1zn , and @xmath2zn have been measured with the activation technique in a quasi - stellar neutron spectrum corresponding to a thermal energy of @xmath3 kev . by a series of repeated irradiations with different experimental conditions , an uncertainty of 3% could be achieved for the @xmath0zn(@xmath4)@xmath5zn cross section and for the partial cross section @xmath1zn(@xmath4)@xmath6zn@xmath7 feeding the isomeric state in @xmath6zn . for the partial cross sections @xmath2zn(@xmath4)@xmath8zn@xmath7 and @xmath2zn(@xmath4)@xmath8zn@xmath9 , which had not been measured so far , uncertainties of only 16% and 6% could be reached because of limited counting statistics and decay intensities .
compared to previous measurements on @xmath10zn , the uncertainties could be significantly improved , while the @xmath2zn cross section was found two times smaller than existing model calculations . from these results
maxwellian average cross sections were determined between 5 and 100 kev . additionally , the @xmath11-decay half - life of @xmath8zn@xmath7 could be determined with significantly improved accuracy .
the consequences of these data have been studied by network calculations for convective core he burning and convective shell c burning in massive stars . |
local helioseismic diagnostic methods such as time - distance helioseismology @xcite , helioseismic holography @xcite and ring - diagram analysis @xcite , have over the years provided us with unprecedented views of the structures and flows under sunspots and active regions . however , a growing body of evidence appears to suggest that interpretations of the measured statistical changes in the properties of the wave - field may be rendered inaccurate by complexities associated with the observations and wave propagation physics . incorporating the full mhd physics and understanding the contributions of phase and frequency filters , and differences in the line formation height ,
are thought to be central to future models of sunspots .
one of the earliest studies that highlighted the interaction of waves with sunspots was the fourier - hankel analysis of @xcite , who found that sunspots can absorb up to half of the incident acoustic - wave power and shift the phases of interacting waves quite significantly ( see also @xcite ) .
these results were echoed over the years by a steady steam of theoretical results ( e.g. @xcite ; @xcite ; @xcite ; @xcite ; @xcite ) that have consistently emphasized the need for more sophisticated modeling and interpretation of wave propagation in strongly magnetized regions .
important advances in our observational understanding of sunspots were also achieved by @xcite and @xcite , who inferred the presence of flows underneath sunspots , and @xcite who estimated the sub - surface wave - speed topology . however , while the inversion procedures applied to derive these results fail to directly account for the tensorial nature of magnetic field effects , the action of the field is mimicked via changes in the acoustic properties of the medium ( the so - called wave speed ) .
recently however , numerical forward models of helioseismic wave ( e.g. @xcite ; @xcite ) and ray @xcite propagation in magnetized atmospheres have been developed and are beginning to make inroads into this problem .
in particular , the results of @xcite and cameron ( 2008 ; private communication ) strongly suggest that active - region magnetic fields play a substantial role in influencing the wave field , and that the complex interaction of magnetic fields with solar oscillations , as opposed changes in the wave - speed , are the major causes of observed travel - time inhomogeneities in sunspots .
we study the impact of strong magnetic fields on wave propagation and the consequences for time - distance helioseismology using two numerical forward models , a 3d ideal mhd solver and mhd ray theory .
the simulated data cubes are analyzed using the traditional surface - focused center - to - annulus method frequently applied in the time - distance analyses of sunspots ( e.g. , @xcite ) .
furthermore , we apply the same method as @xcite to also isolate and analyze the thermal contribution to the observed travel - time shifts .
the background stratification is given by a convectively stable , hydrostatic @xmath1 polytrope below the photosphere , smoothly connected to an isothermal atmosphere .
the magneto - hydrostatic ( mhs ) sunspot model that we embed in the background is similar in construction to that of @xcite and @xcite , where the flux tube is modeled by an axisymmetric magnetic field geometry based on the @xcite self - similar solution .
this approximation requires the following choices for the radial ( @xmath2 ) and vertical ( @xmath3 ) components of the magnetic field : @xmath4 where @xmath5 and @xmath6 are the radial and vertical coordinates in cylindrical geometry .
the term @xmath7 controls the magnitude of the magnetic field ( peak field strength of 3000 g at the photosphere ) and the flux ( @xmath8 ) .
the horizontal extent of the tube and its rate of divergence with altitude is set by @xmath9 ( see figure [ fig : mhd ] a ) . upon solving the mhs equations of pressure and lorentz support ( described in detail in @xcite ) ,
we obtain the altered thermodynamic stratification of the underlying magnetized plasma ( figure [ fig : mhd ] c ) .
the linearized ideal mhd wave equations are integrated according to the recipe of @xcite .
we implement periodic horizontal boundaries and place damping sponges adjacent to the vertical boundaries to enhance the absorption and transmission of outgoing waves .
waves are excited via a pre - computed deterministic source function that acts on the vertical momentum equation . in a manner similar to @xcite ,
the forcing term is also multiplied by a spatial function that mutes source activity in a circular region of 10 mm radius to simulate the suppression of granulation related wave sources in a sunspot .
figure [ fig : mhd ] shows some of the resulting properties of the simulated vertical ( doppler ) velocity data extracted from the mhd simulations .
@xcite detail the steps involved in using mhd ray theory to model helioseismic ray propagation in magnetized atmospheres . here , we provide a brief description of the magneto - acoustic ray tracing procedure .
the ray paths are computed in cartesian geometry in the vertical @xmath10-@xmath6 plane assumed to contain both magnetic field lines and ray paths . in this case , we only require the 2d dispersion relation with the alfvn wave factored out : @xmath11 where @xmath12 , @xmath13 represents the sound speed , @xmath14 the alfven speed , @xmath15 is the squared brunt - visl frequency and @xmath16 is the square of the isothermal acoustic cut - off frequency .
the remaining term , @xmath17 ( where @xmath18 is the prescribed magnetic field ) , represents the component of the wavevector @xmath19 parallel to the magnetic field .
the full 3d dispersion relation is presented in @xcite .
the construction of @xmath19 is completed by specifying the governing equations of the ray paths using the zeroth order eikonal approximation @xcite which are solved using a fourth - order runge - kutta numerical method .
the magneto - acoustic rays stay on the fast - wave dispersion branch at all times .
it should be noted that neither forward model ( i.e. , @xmath20[mhd.hanasoge ] , [ mhd.moradi ] ) accounts for the presence of flows .
the simulated doppler velocity data - cube of @xmath20[mhd.hanasoge ] ( extracted at an observational height of 200 km above the photosphere ) had dimensions of @xmath21 mm@xmath22 @xmath23 @xmath24 minutes , with a spatial resolution of @xmath25 mm and a cadence of 1 minute . for the time - distance calculations , we compute cross covariances of oscillation signals at pairs of points on the photosphere ( source at @xmath26 , receiver at @xmath27 ) based on a single - skip center - to - annulus geometry ( see e.g. @xcite ) .
we cross correlate the signal at a central point with signals averaged over an annulus of radius @xmath28 around that center .
firstly , we filter out the @xmath29-mode ridge . subsequently ,
standard gaussian phase - speed in conjunction with gaussian frequency filters centered at 3.5 , 4.0 and 5.0 mhz with 0.5 mhz band - widths are applied in order to study frequency dependencies of travel times ( e.g. @xcite ; @xcite ) .
the annular sizes and phase - speed filter parameters used in estimating the times shown in figures [ fig : dtmaps ] and [ fig : azim ] ( including the central phase speed ( @xmath30 ) and full width at half - maximum ( fwhm ) used ) are outlined in table [ tab : filters ] . the point - to - annulus cross - covariances
are fitted by two gabor wavelets ( e.g. , @xcite ) to extract the required travel times . in order to compare theory with simulation
, we estimate centre - to - annulus mean time shifts , @xmath31 , using the mhd ray tracing technique of @xmath20[mhd.moradi ] for the same sunspot model ( @xmath20[sunspot ] ) .
the single - skip magneto - acoustic rays do not require filtering .
instead , they are propagated from the upper turning point of their trajectories , in both the positive and negative @xmath10 directions , at a prescribed frequency with horizontal increments of 1 mm across the sunspot .
the required range of horizontal skip distances are obtained by altering the shooting angle at which the rays are initiated .
the skip distances are then binned according to their travel path lengths , @xmath32 , while the travel times are averaged across both the positive and negative horizontal directions . for both forward models (
i.e. , @xmath20[mhd.hanasoge ] , [ mhd.moradi ] ) , we only concern ourselves with the mean _ phase _ time shifts .
figure [ fig : dtmaps ] shows maps of @xmath31 as well as the frequency filtered azimuthal averages of @xmath31 obtained using time - distance center - to - annulus measurements for the measurement geometries indicated in table [ tab : filters ] .
the @xmath31 map for @xmath33 mm clearly displays positive travel - time shifts , reaching a maximum of around 25 seconds at spot center .
a similar travel - time shift is observed from the azimuthal average of @xmath31 when a frequency filter centered at 3.5 mhz is applied to the data .
we also observe the magnitude of the positive @xmath31 steadily decrease as we increase the frequency filter to 4.0 mhz , with negative @xmath31 starting to appear in the profile , and by 5.0 mhz the travel times observed inside the sunspot are completely negative . for the larger annuli , negative time shifts of increasing magnitude are consistently observed as we increase the central frequency of the filter .
in fact , all @xmath31 maps for @xmath32 larger than @xmath34 mm that we measured displayed similar @xmath31 behavior to the @xmath35 and @xmath34 mm bins ( albeit with smaller time shifts ) . it is important to take note of both the signs of the travel - time perturbations and their apparent frequency dependence .
positive @xmath31 have traditionally been interpreted as indicative of a region of slower wave propagation in the shallow subsurface layers beneath the spot , while negative times are of a wave - speed enhancement .
so in essence , the @xmath31 profiles that we have derived from the simulation would appear to indicate a traditional two - layered " wave - speed structure ( e.g. , @xcite ; @xcite ) beneath the sunspot .
however as can be seen in figure [ fig : mhd ] , the thermal profile of our model atmosphere is a `` one - layer '' sunspot model ( @xmath36 ) and of the order of @xmath37 .
similarly , changes in the sub - surface wave speed , @xmath38 ( where @xmath39 is the unperturbed sound speed ) , lie only in the positives @xmath40 0250% ( not shown here ) , with the greatest enhancements seen near the surface .
the large decrease in the sound speed we observe in our model also raises the possibility that current methods of linear inversion may lie beyond their domains of applicability .
we also observed that travel times associated with the smallest measurement geometry are most sensitive to the phase - speed filter used , i.e. , when the phase - speed parameters were adjusted to filter all background power below the @xmath0 ridge , negative @xmath31 were obtained .
this behavior was noted by @xcite , who determined the causative factor to be the background power between the @xmath0 and @xmath29 ridges .
it is unsettling that the sign of the time shift may be reversed at will , through small changes in the filter width and center .
figure [ fig : azim ] ( frames a - c ) show the resultant @xmath31 profiles derived from the mhd ray tracer for identical measurement geometries as used for the time - distance calculations .
the similarities between between the ray @xmath31 profiles and their time - distance counterparts in figure [ fig : dtmaps ] are striking .
firstly , the ray travel - time perturbation profiles contain predominantly negative travel - time shifts for all frequencies , albeit with slightly smaller magnitudes .
secondly , a similar frequency dependence of @xmath31 is also observed .
generally , high frequency rays propagated within the confines of a magnetic field are expected to i ) travel faster and ii ) propagate longer distances than low frequency rays @xcite .
however , one significant difference we can observe in these profiles is the absence of any positive travel - time shifts for the @xmath41 mm bin .
this is significant because the exclusively negative @xmath31 we observe across all geometries not only reflects the one - layered wave- and sound - speed profiles below the surface , but also highlights the effects that phase - speed filtering can have on time - distance measurements ( recall that ray calculations require no such filtering ) .
nonetheless , the overall self - consistency between these results and those in @xmath20[mhdtt ] are very encouraging , despite the 2d nature of the ray calculations . given the fact that ray theory appears to succeed in capturing the essence of the travel - time variations as derived from the mhd simulations , we can isolate the thermal component of the measured @xmath31 using the same approach as @xcite to ascertain the contribution to the travel - time shifts from the underlying thermal structure . to do this
, we essentially re - calculate the ray paths in the absence of the flux tube while maintaining the modified sound - speed profile obtained in @xmath20[sunspot ] .
the resulting _ thermal _ travel - time perturbations , @xmath42 , would then be purely a result of thermal ( sound - speed ) variations along the ray path .
the resulting @xmath42 profiles , presented in figure [ fig : azim ] ( frames d - f ) , surprisingly show that , even without the magnetic field , ray theory produces negative travel times the exception being for rays propagated at 5.0 mhz .
this indicates that the contribution from the underlying thermal structure is significant enough to modify the upper turning point of the ray paths , thus shortening the ray travel times .
the appearance of negative travel times for a model with a decrease in sound speed would appear to be somewhat counterintuitive , since from standard ray theory , one would expect negative time shifts with increases in sound speed .
the most likely explanation for this phenomenon , is that since both the sound speed and density differ from the quiet sun , consequent changes in the acoustic cut - off frequency ( @xmath43 , where @xmath44 is the density scale height ) in the near - surface regions of our model modifies the the ray path for waves with frequencies less than 5.0 mhz quite significantly , thereby causing negative travel - time shifts .
however , when comparing with the time perturbations derived from calculations that include the magnetic field ( i.e. figure [ fig : azim ] frames a - c ) , mhd effects appear to be dominant contributors to the observed time shifts .
this is perhaps most evident for @xmath45 mm ( figure [ fig : azim ] d ) , where the thermal contribution at spot center appears to make up approximately 11% of @xmath31 ( i.e. figure [ fig : azim ] a ) at 5.0 mhz , 18% at 4.0 mhz and 28% at 3.5 mhz . for the largest bin ,
we see a similar contribution at 4.0 and 5.0 mhz , but a much greater contribution at 3.5 mhz ( 45% of @xmath31 ) .
incorporating the full mhd physics into the various forward models used in local helioseismology is essential for testing inferences made in regions of strong magnetic fields . by comparing numerical simulations of mhd wave - field and ray propagation in a model sunspot
, we find that : i ) the observed travel - time shifts in the vicinity of sunspots are strongly determined by mhd physics , although sub - surface thermal variations also appear to affect ray timings by modifying the acoustic cut - off frequency , ii ) the time - distance travel - time shifts are strongly dependent on frequency , phase speed filter parameters and the background power below the @xmath0 ridge , and finally iii ) mhd ray theory succeeds in capturing the essence of center - to - annulus travel - time variations as derived from the mhd simulations .
the most unsettling aspect about this analysis is that despite using a background stratification that differs substantially from model s ( christensen - dalsgaard et al .
1996 ) and a flux tube that clearly lacks a penumbra , the time shifts still look remarkably similar ( at least qualitatively ) to observational time - distance analyses of sunspots .
preliminary tests conducted with different sunspot models ( e.g. , different field configurations , peak field strengths etc . )
have also provided similar results .
given the self - consistency of these results , as derived from both forward models , it could imply that we are pushing current techniques of local helioseismology to their very limits .
it would appear that accurate inferences of the internal constitution of sunspots await a clever combination of forward modeling , observations , and a further development of techniques of statistical wave - field analysis . | we investigate the direct contribution of strong , sunspot - like magnetic fields to helioseismic wave travel - time shifts via two numerical forward models , a 3d ideal mhd solver and mhd ray theory .
the simulated data cubes are analyzed using the traditional time - distance center - to - annulus measurement technique .
we also isolate and analyze the direct contribution from purely thermal perturbations to the observed travel - time shifts , confirming some existing ideas and bring forth new ones : ( i ) that the observed travel - time shifts in the vicinity of sunspots are largely governed by mhd physics , ( ii ) the travel - time shifts are sensitively dependent on frequency and phase - speed filter parameters and the background power below the @xmath0 ridge , and finally , ( iii ) despite its seeming limitations , ray theory succeeds in capturing the essence of the travel - time variations as derived from the mhd simulations . |
renormalization group ( rg ) flow describes the change in the effective behaviour of a quantum field theory as the distance or energy scale is changed . in the wilsonian scheme one starts from a bare theory in the ultraviolet ( uv ) and integrates out high energy degrees of freedom .
the effect of integrating out the high energy degrees of freedom is to redefine or renormalize the bare coupling constants appearing in the lagrangian of the uv theory .
this redefinition is mathematically described as a first order flow equation in the space of coupling constants .
conformal field theories ( cft ) appear as fixed points of this flow .
now given a cft ( uv - cft ) one can perturb this by adding a relevant operator and flow to a new cft ( ir - cft ) in the infrared ( ir ) .
for two dimensional unitary and lorentz invariant quantum field theories it has been shown in @xcite that one can define a positive monotonically decreasing function of the coupling constants along the flow whose values at the fixed points coincide with the central charges of the respective two dimensional cfts .
this positive monotonically decreasing function of the coupling constants is known as a @xmath0-function .
@xmath0-theorem was generalized and proved for higher dimensional field theories in @xcite .
rg flow can also be studied via ads / cft correspondence . in this
set up one usually starts with a bulk theory which has several pure ads critical points .
the bulk theory is usually a classical theory of gravity coupled to matter fields .
rg flow is represented by a domain - wall like solution of the coupled gravity - matter system which interpolates between two ads critical points @xcite.the holographic version of the @xmath0-theorem is proved by showing the existence of a function which decreases monotonically from the uv - ads to the ir - ads as a result of the field equations and gives the central charges when evaluated at the end points of the flow . in all these constructions
null energy condition on the matter stress tensor plays an important role .
the story of the holographic @xmath0-theorem is more interesting when one considers higher derivative corrections to einstein gravity in the bulk .
this was considered in great detail in @xcite .
higher derivative corrections are ubiquitous in low energy effective theories coming from string theory and they are also interesting in their own right as toy models where one can study various thorny issues related to black hole thermodynamics . in an arbitrary theory of gravity described by a local covariant action there is in general be no clear separation between gravity and matter sector .
so matter stress tensor can not be uniquely defined and standard null energy condition has no clear role to play .
this is one of the main obstacles of proving holographic @xmath0-theorem in a general theory as was emphasized in @xcite .
the other point is that even if there is a higher derivative theory of gravity coupled to matter where null energy condition can be imposed unambiguously , sometimes it is not enough to prove holographic @xmath0-theorem and validity of the @xmath0-theorem seems to require additional constraints on the pure gravity sector of the theory @xcite .
this could be a genuine physical constraint on the bulk theory like unitarity but its status is always not so clear @xcite .
so an important problem is to spell out the constraint which needs to be imposed on a theory of gravity so that holographic @xmath0-theorem holds . in @xcite , a different approach towards holographic rg flow and holographic @xmath0-theorem
was proposed .
the basic observation of @xcite was that causal horizons in pure poincare ads are killing horizons generated by the dilatation vector .
dilatation symmetry is broken along the rg flow and as a result causal horizons in a rg flow geometry are no longer globally killing .
they become killing ( stationary ) in the asymptotic uv and ir region of the rg flow geometry , where scaling is recovered . in @xcite
it was shown that the bekenstein - hawking entropy ( density ) of the non - killing causal horizon interpolating between the uv and the ir of the rg flow geometry is a holographic @xmath0-function .
the monotonicity of the @xmath0-function followed from the second law of thermodynamics obeyed by causal horizons in ads @xcite .
based on these observations it was proposed in @xcite that the boundary rg flow is dual to the thermodynamics of causal horizons in ads and the gravity dual of the @xmath0-theorem is the second law of causal horizon thermodynamics . in this paper we verify this for higher derivative theories of gravity which admit pure @xmath1 as a solution . at
first we test it for higher derivative theories for which an entropy expression exists which satisfies at least the linearized version of the second law .
it has been shown recently @xcite that a general curvature - squared theory of gravity satisfies the linearized second law .
we verify that in these cases the entropy ( density ) evaluated on the spatial slices of the dynamical causal horizon is a holographic @xmath0-function and the gravity dual of the boundary @xmath0-theorem is the linearized second law .
this restricts us to rg flows where the difference between the uv and the ir central charges is small , but there is no restriction on the magnitude of the higher - derivative coupling . in all these cases
the bulk gravity theory is minimally coupled to the matter sector which cause the rg - flow and the matter stress tensor is restricted to satisfy the null - energy condition .
we would like to emphasize that this does _ not _ mean that we restrict to gravity - matter theories which satisfy the null convergence condition for any future directed null vector @xmath2 , where @xmath3 is the ricci tensor .
if this condition is satisfied then raychaudhuri equation applied to a hypersurface orthogonal congruence of null geodesics implies that the expansion @xmath4 is monotonically decreasing along the generators of the congruence .
when einstein gravity is minimally coupled to the matter field the null convergence condition is guaranteed if the matter stress tensor satisfies the null energy condition .
however this is not the case if there are higher derivative couplings .
even if the matter field is minimally coupled and the matter stress tensor satisfies the null energy condition , this does not guarantee the null convergence condition due to the presence of the higher derivative terms .
so focusing does not happen and the proof of the second law can not follow the same logic as in the einstein gravity case . ] which is generically violated in all higher - derivative theories of gravity including the ones we are studying .
our next example is @xmath5 gravity where @xmath6 is the ricci scalar . in this case
the entropy expression satisfying the second law to all orders is known @xcite and we construct the holographic @xmath0-function using this .
this is an example of a theory which has fourth order equation of motion , but the bulk theory is still unitary @xcite .
we also study the effect of non - minimal coupling to the matter sector . as an example
, we treat the case of einstein gravity plus scalar matter where in addition to the standard minimal coupling there is also a curvature coupling of the form @xmath7 , where @xmath6 is the ricci scalar . in this case
there is no clean separation between matter and gravity sector and as a result standard null energy condition can not be defined . in such a theory
the second law is known to hold , albeit with an expression which is not the area of a space - like slice of the horizon @xcite . using this expression we prove holographic @xmath0-theorem in a situation where the null - energy condition can not be defined . as a generalization of this
we show that a sufficient condition to ensure holographic @xmath0-theorem in a theory of gravity described by a local covariant action in ads is that the second law of causal horizon thermodynamics be satisfied by the theory .
of the boundary point p in pure ads .
this is just the bulk light - cone of the point p. poincare time runs vertically upward in the diagram.,width=566 ] in this section we will briefly review the contents of @xcite and discuss some notations . let us consider @xmath8 written in poincare coordinates , ds^2=(-dt^2 + _
i=1^d-1(dx^i)^2 + dz^2 ) where @xmath9 is the @xmath1 radius .
it will be useful to introduce the spherical polar coordinates such that , _
i=1^d-1 ( dx^i)^2 = dr^2 + r^2d_d-2 ^ 2 where , @xmath10 is the metric of a unit @xmath11-sphere . in these coordinates , @xmath1-metric is , ds^2=(-dt^2 + dr^2 + r^2d_d-2 ^ 2 + dz^2 ) .
this metric has a scaling isometry , @xmath12 , generated by the dilatation vector , d=+ + @xmath13 is null on the hypersurface @xmath14 given by , -t^2 + r^2 + z^2 = 0 @xmath14 is a null hypersurface which is left invariant by the flow of @xmath13 and so @xmath14 is a killing horizon generated by @xmath13 .
it is also easy to see that @xmath14 is the boundary of the causal past ( future ) of the point , @xmath15 , in @xmath1 .
so @xmath14 is a causal horizon and in pure @xmath1 it is a killing horizon generated by the dilatation vector , @xmath13 .
let us now deform the boundary field theory by adding a relevant operator which is dual to a scalar field of negative mass squared in @xmath1 .
this induces rg - flow . for simplicity
we will take the backreacted geometry to be , [ domain ] ds^2=(-dt^2 + dr^2 + r^2d_d-2 ^ 2 + ) where @xmath16 as @xmath17 and @xmath18 as @xmath19 . here
@xmath20 is the radius of the locally @xmath1 region which develops near the poincare horizon , @xmath21 .
this metric has no global scaling isometry because dilatation is broken along the rg - flow .
scaling is recovered only asymptotically as @xmath17 or @xmath19 .
let us now write the metric in the form , [ remetric ] ds^2=(-dt^2 + dr^2 + r^2d_2 ^ 2 + d|z^2 ) where we have defined , [ barz ] = _ 0^z = _ 0 ^ 1 in ( [ remetric ] ) , @xmath22 should be treated as a function of @xmath23 given implicitly by the relation ( [ barz ] ) . throughout the paper prime over @xmath22 will denote differentiation with respect to @xmath23 .
the causal horizon of the point , @xmath15 , in the back - reacted geometry can now be written as , [ dcausal ] -t^2 + r^2 + |z^2 = 0 of the boundary point p in deformed ads .
this is just the deformed bulk light - cone of the point p. poincare time runs vertically upward in the diagram.,width=491 ] let us denote this by @xmath24 .
@xmath24 is again a null - hypersurface and we can take as its null generator the vector field @xmath25 given by , @xcite @xmath25 is null on @xmath24 .
@xmath25 is not a killing vector .
it coincides with @xmath13 only asymptotically where scaling is recovered . as a result
the backreacted causal horizon @xmath24 is * no * longer a killing horizon .
it is dynamical in the sense that it has non - zero expansion and shear .
in @xcite it was proposed that the boundary rg - flow is dual to the ( thermo ) dynamics of causal horizons caused by the breaking of the scaling isometry in the bulk . in particular
it was shown in @xcite that if the bulk theory is described by einstein - hilbert action minimally coupled to a scalar field obeying the null - energy condition then the gravity dual of the boundary @xmath0-theorem is the second law of thermodynamics obeyed by causal horizons @xcite .
the holographic @xmath0-function in this framework was given by the bekenstein - hawking entropy density of the dynamical causal horizon ( [ dcausal ] ) .
_ this directly relates the second law for causal horizons in @xmath1 to the unitarity of the boundary theory in the form of @xmath0-theorem_.
we first consider the einstein - gauss - bonnet theory in @xmath26 given by the action , i_gb =- d^5x ( r++(r_r^-4r_r^+r^2 ) ) .
if we minimally couple it to scalar matter which induces the rg - flow , then the equation of motion for this theory is given by , [ met3 ] r_- r g_- g_- h_=t _ where , @xmath27 and @xmath28 is the matter stress tensor .
we have assumed that the matter stress - tensor obeys the null - energy condition .
so if we contract this with the null vector @xmath25 we will get , [ met2 ] r_|d^|d^-h_|d^|d^0 . evaluating ( [ met2 ] ) on ( [ remetric ] ) we get the following condition , [ met4 ] 0 . where @xmath29 .
we will analyze this condition later .
let us now turn our attention to the entropy functional which is the jacobson - myers ( jm ) one in this case @xcite s_jm = d^3 x ( 1 + l^2 ) .
the integral is taken along a space - like slice of the horizon ( [ dcausal ] ) and @xmath30 is the intrinsic ricci scalar of the space - like slice .
jm entropy functional satisfies the linearized second law . to evaluate this
, it is useful to parametrize the causal horizon ( [ dcausal ] ) as , [ parahyp ] t = |z , r=|z , |z=|z this parametrization corresponds to a particular space - like slicing of the causal horizon . with this parametrization
the induced metric on the space - like slices of ( [ dcausal ] ) becomes , .
] [ hyp ] ds_ind^2= l_ads^2 ( ) ^2 ( d^2 + ^2 d_2 ^ 2 ) = l_ads^2 ( ) ^2 ds_h^3 ^ 2 where @xmath31 is the metric on a unit hyperboloid .
so , r = - ( ) ^2 now following @xcite we take the null generators of the causal horizon as integral curves of @xmath25 defined as , [ inte ] = t , = r , = |z , where @xmath32 is not necessarily an affine parameter .
the triple @xmath33 satisfies the condition , [ nullcondition ] -t()^2 + r()^2 + |z()^2 = 0 it follows from ( [ parahyp ] ) and ( [ inte ] ) that @xmath34 is a comoving coordinate , i.e @xmath34 remains constant along the integral curves satisfying ( [ inte ] ) and ( [ nullcondition ] ) and so all the coordinates parametrizing the unit hyperbolic space are comoving .
let us now define the density of jm entropy of the causal horizon as , ds_jm = d^3 x ( 1 + l^2 ) = 2 ( ) ^3 ( 1- 6 ( ) ^2 ) dv_h^3 since the volume of the unit hyperboloid is independent of the parameter @xmath32 of the null generator , second law becomes a statement about the evolution of the finite coefficient along the null generator .
we denote the finite coefficient by @xmath35 , [ gbc ] a ( ) = 2 ( ) ^3 ( 1- 6 ( ) ^2 ) = 2 ( ) ^3 ( 1- 6f _ ( ) ^2 ) this gives the correct @xmath36-central charge at the uv ( @xmath37 ) and ir ( @xmath38 ) ) that as @xmath37 , @xmath39 and as @xmath40 , @xmath41 .
it follows from a simple argument of @xcite that @xmath17 as @xmath39 and @xmath19 as @xmath42 . so as @xmath32 runs from @xmath43 to @xmath44 , the null generator of the causal horizon runs from the uv to the ir . ]
fixed points . to see this
more quantitatively @xcite , we note that as @xmath37 , @xmath17 and @xmath45 .
similarly as @xmath40 , @xmath21 and @xmath46 . substituting this we get the correct values of the uv and the ir central charges , respectively .
function following the references of @xcite where the expression for the entropy and the central charge have been worked out . ]
the monotonicity follows from the linearized second law satisfied by the jm entropy function , but we will check this by explicit calculation . since the jm entropy satisfies
the linearized second law , the @xmath0-function ( [ gbc ] ) constructed out of the jm entropy function is monotonic if we confine ourselves to linearized metric fluctuations . by linearized we mean that the metric fluctuations around the uv - ads is of the same order of magnitude as the matter stress tensor causing the rg flow .
this condition allows us to probe holographic rg flows where the difference between the uv and the ir central charges is small .
this does * not * mean that gauss - bonnet coupling constant @xmath47 is small .
now , [ avariation ] =3 ( ) ( 1-_0 ^ 1 ) ( ( _ 0 ^ 1)^2 - 2f _ ) .
we want to show that , @xmath48 one can check that this condition can not be proved in complete generality without imposing an extra condition that @xmath49 . but this does not follow from the null energy condition , \0 . since @xmath50 is positive
this condition is equivalent to , [ ne ] f(z ) ( 1 - 2 f _
f(z))0 .
the fact that we can not prove ( [ monotony ] ) in complete generality using the condition ( [ ne ] ) is just the reflection of the fact that the jm entropy functional satisfies only the linearized second law .
so let us do a linearized calculation .
let us assume that the matter stress tensor is of order @xmath51 . since in the fixed point geometry @xmath52 is a constant , the leading term of @xmath53 is of order @xmath51 and linearized calculation means we keep only order @xmath51 terms in our calculation .
so in this approximation , the null - energy condition becomes , [ lne ] f(z ) ( 1 - 2 f _ ) 0 . applying the linearization to ( [ avariation ] ) we get , =3 ( 1-_0 ^ 1 ) ( 1 - 2f _ ) .
now we can write , 1-_0 ^ 1 = - _ 0
^ 1 _ z^z f ( ) using this we can write , [ savariation ] = - _ 0 ^ 1 _ z^z f ( ) ( 1 - 2f _ ) .
the expression on the rhs of ( [ savariation ] ) is manifestly negative once we apply the linearized null - energy condition ( [ lne ] ) .
so , 0 to linearized order .
so the @xmath0-function constructed from jm entropy functional decreases monotonically as a consequence of the linearized second law for causal horizons .
this is the best we can hope from the jm entropy functional . we now consider the general curvature squared theory in @xmath26 given by the action , i_r^2=-d^5x ( r++(_1r_r^+_2 r_r^+_3 r^2 ) ) .
we minimally couple it to scalar matter which induces the rg - flow .
an entropy functional which satisfies the second law to all orders in this theory is not known .
however , an entropy functional which satisfies the linearized second law has been written down recently in @xcite . in order to write it down
we define another auxiliary vector @xmath54 such that @xmath55 and @xmath56 on the horizon ( [ dcausal ] ) .
also @xmath54 is orthogonal to the spacelike slice of the horizon .
|z ) with this the entropy functional for this theory can be written as @xcite , [ entr2 ] s = d^3 x(1 + l^2 ( _ 3r-4_2 r
_ |d^l^+_1r_|d^ l^ |d^ l^+ ( 4 _ 1 + 3 _ 2)_|d_l ) ) . where @xmath57 and @xmath58 are respectively the two expansion along @xmath25 and @xmath59 given by , _ |d=- and _
l=. in general the shear term also enter in the entropy functional , but in our case it vanishes .
we get the @xmath0-function after evaluating ( [ entr2 ] ) on ( [ hyp ] ) , [ c2 ] a()=2()^3 ( ) . where @xmath60 and @xmath61 .
it gives correct central charges at the fixed points .
a=2(1 - 2 f_(_1 + 2(_2 + 5_3 ) ) ) .
it has been shown in @xcite that ( [ entr2 ] ) obeys linearized second law and as a consequence it follows that @xmath35 is monotonic .
in this section we will consider the example of @xmath5 gravity which is an example of higher curvature theory for which the entropy expression satisfying the second law to all orders is known @xcite .
this is also a theory whose the equation of motion is fourth order , but it still gives rise to a _
unitary _ bulk theory @xcite
. we will construct the corresponding @xmath0 function in this section .
the action in 5 dimensions is , i =- d^5x . where @xmath6 is the ricci scalar .
this action is minimally coupled to scalar field whose energy - momentum tensor satisfies the null energy condition and this induces the holographic rg flow .
the corresponding entropy functional which satisfies the second law is given by @xcite , s = d^3x where the prime denotes derivative with respect to the ricci scalar @xmath6 .
the integral is over the space like slice ( [ parahyp ] ) of the causal horizon ( [ dcausal ] ) . since the ricci scalar for the domain wall geometry ( [ domain ] ) depends only on @xmath22 ( or @xmath23 )
, we can repeat the same arguments as in the previous sections and immediately write down the expression of the @xmath0-function as , [ af ] a()=()^3(1+f(r ( ) ) ) . where r()= and @xmath62 .
in this expression @xmath63 is given implicitly by the equation following from ( [ barz ] ) , = _ 0 ^ 1 and @xmath64 satisfies ( [ inte ] ) .
so @xmath65 is the ricci scalar of the domain wall geometry evaluated on the null generator of the causal horizon .
the monotonicity ( @xmath66 ) of this @xmath0 function now follows from the proof of the second law in @xcite .
let us now make a few comments on the formula ( [ af ] ) .
the proof of the second law in @xcite in case of @xmath5 gravity relied on the crucial assumption , @xmath67 , besides null energy condition on matter stress tensor .
we can see that in the context of holographic rg flow this positivity condition gets translated into the * unitarity * of the dual boundary theory .
if the positivity condition is violated then the @xmath0-function ( [ af ] ) can be negative which signals non - unitarity . in our case
it is the _ entropy density _ of the dynamical causal horizon ( [ dcausal ] ) which plays the role of holographic @xmath0-function and so the positivity is essential for its interpretation as the number of degrees of freedom of the dual field theory along the rg flow .
so we can see that in the case of causal horizons in ads proof of the second law and unitarity are very closely related .
we would also like to make a comparison between our formula for the @xmath0-function ( [ af ] ) and the one proposed in @xcite .
the authors of @xcite proposed a formula for the @xmath0-function in @xmath5 gravity whose monotonicity relied on the extra assumption which in our notation becomes @xmath68 . here
prime denotes derivative with respect to @xmath6 .
this is a condition on the gravitational sector of the theory . in our case
this extra assumption does not play any role .
the monotonicity of the @xmath0-function ( [ af ] ) follows directly from the second law and the second law holds under the assumption of the null energy condition of the matter stress tensor and @xmath67 . so in this sense our @xmath0-function is more universal than that proposed in @xcite .
let us consider the following action in 5 dimensions , [ nmc ] i =- d^5x . where @xmath69 is a scalar field which triggers the rg flow .
the importance of more general coupling between gravity and matter was emphasized in @xcite .
we assume that this action yields several critical points with constant values of the scalar field and the bulk geometry being pure @xmath26 .
we have a non - trivial curvature coupling , @xmath70 , in this action .
the equation of motion of this theory is given by , @xmath71 where @xmath72 is the einstein tensor .
we can see from the equation of motion and also from the action that there is no clear separation between matter and gravity .
in particular there is no unique way of defining matter stress tensor .
for example one could define an effective matter stress tensor by solving for the einstein tensor and defining , _ eff = g _
= this effective matter stress tensor violates null energy condition for any non - zero value of the coupling parameter @xmath73 @xcite . in our framework
this is not a problem as long as we have an entropy function for the gravity - matter system , described by ( [ nmc ] ) , which satisfies the second law . the validity of the second law does not depend on our ability to define a cleanly separated matter sector .
such an entropy function indeed exists for this theory which satisfies the second law to all orders .
it is given by @xcite , [ nonminimal ] s = d^3 x ( 1 - 8 ^ 2 ) where the integral is over a space - like slice of the dynamical horizon .
the form of the entropy function ( [ nonminimal ] ) is valid for any form of the scalar self - interaction @xmath74 .
let us now assume that in the domain wall geometry the scalar field @xmath69 depends only on the coordinate @xmath22(or @xmath23 ) . using this and the same steps as in the previous sections we arrive at the expression of the @xmath0 function to be , [ nonc ] a()=()^3(1 - 8 ^2(z ( ) ) ) it gives the correct central charge at the fixed point . in the above formula
the scalar field @xmath69 is evaluated along the null generator of the causal horizon and so it is a function of the parameter @xmath32 .
the monotonicity of the @xmath0-function ( [ nonc ] ) follows from the second law @xcite .
so the condition that a gravity - matter system satisfies the second law of causal horizon thermodynamics in ads is more fundamental , because it is enough to prove the holographic @xmath0-theorem even in cases where null energy condition on matter stress tensor can not be defined .
hence : _ the condition , which replaces the null energy condition on matter stress tensor , is that the second law of causal horizon thermodynamics @xcite be satisfied by the gravity - matter system in ads .
this also includes theories with higher derivative couplings_. this condition has the virtue that this does not require a separation between the gravity and the matter sector because the validity of the second law does not require any such separation .
we will now argue that this condition is sufficient for proving holographic @xmath0-theorem in an arbitrary gravity theory described by a local covariant action . in order to do that we consider an arbitrary local and diffeomorphism invariant theory in the bulk described by the action , [ hac ] i = d^d+1x ( g_ab , r_abcd , _ e r_abcd, ..........
,matter ) where various quantities have standard meaning .
an action of this form will not in general allow a clean separation between matter and gravity sector and standard null - energy condition has no role to play .
let us now assume that this theory admits several ads critical points with constant values of the matter field which we assume to be scalar for simplicity .
we consider a domain wall geometry of the form ( [ domain ] ) which interpolates between a uv - ads ( @xmath17 ) and an ir - ads ( @xmath21 ) .
this interpolating geometry represents the holographic rg flow @xcite . now to prove the holographic @xmath0-theorem we consider causal horizons in this domain - wall geometry , following @xcite .
let us assume that the theory of gravity described by ( [ hac ] ) satisfies the second law of causal horizon thermodynamics .
this means that there exists an entropy expression which satisfies the second law of thermodynamics for causal horizons and reduces to the * wald entropy * when the causal horizon is killing .
wald entropy for the theory ( [ hac ] ) is given by @xcite , [ wald ] s_w = -2d^d-1y _ ab_cd where the integral is done over a codimension two space - like slice of the killing horizon and @xmath75 is the induced metric .
@xmath76 is the binormal to the horizon slice .
we would like to emphasize that this is not the same entropy expression which satisfies the second law .
we will use this formula later . to proceed
let us further assume that the dynamical entropy is again given by the integral of a local geometric expression on space - like slices of the dynamical ( non - killing ) horizon .
schematically , [ dwald ] s_dyn ( ) = -2 _ [ ] d^d-1y s where @xmath32 is a real parameter , not necessarily affine , along the null generators of the horizon and @xmath77 $ ] denotes the @xmath78 constant space - like slice . according to our assumption @xmath79 reduces to the wald expression ( [ wald ] ) when evaluated on a stationary horizon . since we are considering a dynamical horizon
@xmath80 is a non - constant function of @xmath32 and the statement of the ( local ) second law is , [ dent ] ( -2s ) 0 we are getting a monotonic decrease rather than increase because we are looking at a _
past _ horizon .
let us now apply it to our case .
we follow the same strategy as in @xcite and in the previous sections , except that we are now doing it for an arbitrary theory ( [ hac ] ) obeying the second law .
we take the entropy expression ( [ dent ] ) and evaluate it on the space - like cross - section ( [ parahyp ] ) of the non - killing causal horizon ( [ dcausal ] ) .
@xmath32 is the non - affine parameter , parametrizing the null generators ( [ inte ] ) along which the space - like slice @xmath77 $ ] is evolving . the induced metric @xmath75 on @xmath77 $ ] is given by ( [ hyp ] ) .
now in the domain wall geometry all the geometric quantities and the matter fields depend only on the @xmath22 ( or @xmath23 ) coordinate and so is constant on @xmath77 $ ] . as a result @xmath79 is also constant on @xmath77 $ ] although it has non - trivial dependence on @xmath32 through its dependence on @xmath22 ( or @xmath23 ) .
now for the induced metric @xmath75 we get from ( [ hyp ] ) , d^d-1y = l_ads^d-1 ( ) ^d-1 dv_h^d-1 where @xmath81 is the volume element of a unit hyperbolic space space .
since the intrinsic coordinates parameterizing the hyperbolic space ( or @xmath77 $ ] ) are comoving , @xmath81 is independent of @xmath32 .
so the second law ( @xmath82 ) can be written as , [ monotony ] ( -2l_ads^d-1 ( ) ^d-1 s ) 0 so we define our @xmath0-function to be , [ a ] a ( ) = -2l_ads^d-1 ( ) ^d-1 s ( ) the monotonicity of the @xmath0-function , @xmath83 , follows from ( @xmath84 ) .
now we show that at the uv ( @xmath37 ) and the ir ( @xmath40 ) fixed points the @xmath0-function ( @xmath85 ) gives the correct central charges . ) that as @xmath37 , @xmath39 and as @xmath40 , @xmath41 .
it follows from a simple argument of @xcite that @xmath17 as @xmath39 and @xmath19 as @xmath42 .
so as @xmath32 runs from @xmath43 to @xmath44 , the null generator of the causal horizon runs from the uv to the ir .
] near the uv and the ir fixed points the causal horizon ( [ dcausal ] ) again becomes * killing * ( stationary ) because scaling is recovered .
so to evaluate the central charge at the uv and the ir fixed points we can use the wald entropy formula , i.e , a ( - ) = -2l_ads^d-1 ( ) ^d-1 s()|_- = -2l_ads^d-1 s_wald |_uv - ads and a ( ) = -2l_ads^d-1 ( ) ^d-1 s()| _ = -2l_ads^d-1 ( ) ^d-1 s_wald |_ir - ads where , s_wald = _ ab_cd and we have used the fact that @xmath45 as @xmath37 and @xmath46 as @xmath38 .
the @xmath86 has to be evaluated on the uv - ads and the ir - ads .
this was evaluated in @xcite for a general bulk gravity theory described by the action ( [ hac ] ) .
their result is , the ads radius and everything is evaluated on pure ads . using this and the fact that @xmath87 , we get , s_wald |_ads = _ ab_cd |_ads = |_ads therefore , a ( - ) = -2l_ads^d-1 s_wald |_uv - ads = -2 |_uv - ads and a ( ) = -2l_ads^d-1 ( ) ^d-1 s_wald |_ir - ads = -2
|_ir - ads due to maximal symmetry of pure ads the lagrangian density @xmath88 evaluated on pure ads is just a constant number .
now for even @xmath89 the coefficient of the euler density in the trace - anomaly is given by @xcite , a^(d ) = - |_ads so we can see that in even @xmath89 , a ( - ) = 2 a^(d)_uv and a ( ) = 2 a^(d)_ir so these two equations together with the monotonicity condition ( @xmath84 ) @xmath83 following from the second law , proves the holographic @xmath0-theorem for even @xmath89 .
for odd @xmath89 there is no trace - anomaly .
however , as shown in @xcite , for odd @xmath89 , the number @xmath90 can be interpreted in terms of the entanglement entropy of the dual conformal field theory .
we refer to these references for further details .
the calculations in @xcite and in the previous sections show that whenever the second law of causal horizon thermodynamics holds in a theory of gravity coupled to matter in @xmath1 , the entropy density evaluated on the deformed ( non - killing ) causal horizon ( @xmath91 ) is a holographic @xmath0-function .
this is consistent with the proposal of @xcite which says that the boundary rg flow is dual to the thermodynamics of causal horizons , i.e , the validity of the second law is both a * necessary * and a * sufficient * condition for holographic @xmath0-theorem to hold .
we have proved the sufficiency condition in the last section for a general theory of gravity in the bulk .
this raises an interesting puzzle .
if the dual of the boundary @xmath0-theorem is the second law of thermodynamics satisfied by causal horizons in @xmath1 , then what happens to bulk theories which violate the second law ?
if the bulk theory is @xmath92 then this is not a problem , because the dual boundary theory is also non - unitary and @xmath0-theorem does not necessarily hold .
but if the bulk theory is unitary then the dual field theory is unitary and @xmath0-theorem must hold .
so : _ if the duality between @xmath0-theorem and the second law @xcite is correct , then this implies that every * unitary * bulk theory of gravity coupled to matter , * including those with higher - derivative couplings * , satisfies the * second law of causal horizon thermodynamics * in @xmath1 .
if this is not true then @xmath0-theorem will be violated in a unitary lorentz invariant field theory_. in any case , we hope to have convinced the reader that there is a deep connection between second law of causal horizon thermodynamics and unitarity and lorentz invariance of the holographic dual field theory .
this needs to be fleshed out .
moreover our proof of the sufficiency condition in ( [ suffi ] ) shows that for a unitary bulk theory the _ density _ of the dynamical entropy as defined in ( [ dwald ] ) must be positive , i.e , @xmath93 .
if this condition is violated at any point of the dynamical causal horizon then the holographic @xmath0-function as defined in ( [ a ] ) will be negative .
this negativity is a signature of non - unitarity .
the monotonicity of the @xmath0-function also requires the positivity of the entropy production at every point of the causal horizon .
this is usually satisfied in all known cases but this could be a non - trivial constraint on a general theory of gravity in the bulk . hence _ * unitarity * and lorentz invariance of the holographic dual field theory requires a positive entropy density which is non - decreasing at every point of the dynamical causal horizon_. non - unitary theories can also satisfy the second law , but it is not a necessary consequence of the duality between @xmath0-theorem and the second law .
let us now say a few words about the implication of this for higher - derivative theories of gravity in @xmath1 .
= space of all theories in ads ; @xmath94 = space of all theories which obey the second law of causal horizon thermodynamics in ads ; @xmath95 = space of all * unitary * theories in ads .
, width=491 ] in higher derivative theories the entropy of a ( bifurcate killing horizon ) stationary black holes is given by the wald entropy @xcite which by construction satisfies the first law of thermodynamics .
the story of the second law is different . it is known that the wald entropy does not in general satisfy the second law .
there are some cases @xcite where an entropy function exists which satisfies the second law and reduces to the wald entropy in the stationary situation , but no general expression exists .
in fact it is an open question if every higher derivative theory of gravity satisfies the second law or not .
we have argued that for causal horizons in @xmath1 the second law must be satisfied by every _
unitary _ higher - derivative theory of gravity as a consequence of the duality between @xmath0-theorem and the second law .
this is an indirect holographic argument and uses unitarity crucially .
figure-3 summarizes our claim .
an interesting question is whether @xmath96 or @xmath97 is a proper subset of @xmath98 .
that is whether every classical gravity theory satisfies the second law of causal horizon thermodynamics or not ? by every classical gravity theory we mean every covariant unitary or non - unitary gravity plus matter theory in ads described by an action of the form ( [ hac ] ) .
we do not know if this is true or not , but our results show that if this is true , then : _ every unitary or non - unitary field theory with a classical gravity dual satisfies the @xmath0-theorem . in other words , in the large - n but finite @xmath47 limit , @xmath0-theorem becomes insensitive to the unitarity property of field theory_. here @xmath47 is the t hooft coupling of the dual field theory .
before we conclude we would like to discuss some possible extension of our work .
one such thing is to study the @xmath0-function in a thermal state which is dual to a black hole geometry @xcite .
another potential application is to the study rg flows in the attractor geometry @xcite .
another important problem is to understand the proper field theory interpretation of the results of @xcite and those presented in this paper . in short ,
our results show that at least in the large-@xmath99 limit rg flow and @xmath0-theorem has an emergent thermodynamics interpretation where the @xmath0-function plays the role of the entropy density .
in fact the @xmath0-function at the fixed point has a similar interpretation found in @xcite .
our results show that such an interpretation must exist even away from the fixed point .
it will also be interesting to see how boundary rg flow @xcite fits into this framework .
we hope to return to them in near future .
we thank aninda sinha for useful discussions and collaboration in the early stages of this work .
ab would like to thank sudipta sarkar and srijit bhattacharjee for illuminating discussions .
ab thanks iit - gandhinagar for hospitality during the early stages of this work .
ab also like to thank institute of physics , bhubaneswar where the part of this work is presented in the _ young researchers conference
15_. sb would like to thank jyotirmoy bhattacharya , simeon hellerman , kentaro hori , raghu mahajan , djordje radicevic , steve shenker and masahito yamazaki for very helpful discussions and insightful questions .
sb would also like to thank the string theory group at stanford university ( sitp ) for hospitality during the initial stages of this project .
sb would like to thank shiraz minwalla for an interesting discussion on second law in higher derivative gravity during a visit to tifr , mumbai .
we would also like to thank swagata ghatak for her assistance with the figures .
the work of sb was supported by world premier international research center initiative ( wpi ) , mext , japan .
in this section we will demonstrate the monotonicity of ( [ c2 ] ) for a certain restricted class of curvature squared theories following @xcite .
the constraint can be imposed by setting @xmath100 .
we first write down the equation of motion for this theory .
[ met3a ] r_- r g_- g_- h_=t _ where , @xmath101 and @xmath28 is the matter stress tensor .
we have assumed that the matter stress - tensor obeys the null - energy condition .
so if we contract this with the null vector @xmath25 we get , [ met2a ] r_|d^|d^-h_|d^|d^0 . evaluating ( [ met2a ] ) on ( [ remetric ] ) we get the following condition after evaluating on the horiozn @xmath102 , @xmath103 where , @xmath104 when @xmath105 the radial equation of motion simplifies and all the terms containing more than two derivatives of @xmath22 cancel @xcite .
so setting @xmath100 we get , @xmath106 then applying the lineariztion to ( [ met4a ] ) we get , f(z)(1-f_(_2 + 8_3)f(z))0 .
j. l. cardy , `` is there a c theorem in four - dimensions ?
, '' phys .
b * 215 * ( 1988 ) 749 .
+ h. osborn , `` derivation of a four - dimensional @xmath0 theorem , '' phys .
b * 222 * ( 1989 ) 97 .
+ z. komargodski and a. schwimmer , `` on renormalization group flows in four dimensions , '' jhep * 1112 * ( 2011 ) 099 [ arxiv:1107.3987 [ hep - th ] ] .
+ z. komargodski , `` the constraints of conformal symmetry on rg flows , '' jhep * 1207 * , 069 ( 2012 ) [ arxiv:1112.4538 [ hep - th ] ] .
+ h. casini and m. huerta , `` a finite entanglement entropy and the c - theorem , '' phys .
b * 600 * ( 2004 ) 142 [ hep - th/0405111 ]
. + h. casini and m. huerta , `` on the rg running of the entanglement entropy of a circle , '' phys .
d * 85 * , 125016 ( 2012 ) [ arxiv:1202.5650 [ hep - th ] ] .
+ d. l. jafferis , i. r. klebanov , s. s. pufu and b. r. safdi , `` towards the f - theorem : n=2 field theories on the three - sphere , '' jhep * 1106 * , 102 ( 2011 ) [ arxiv:1103.1181 [ hep - th ] ] .
r. c. myers and a. sinha , `` holographic c - theorems in arbitrary dimensions , '' jhep * 1101 * , 125 ( 2011 ) [ arxiv:1011.5819 [ hep - th ] ] .
r. c. myers and a. sinha , `` seeing a c - theorem with holography , '' phys .
d * 82 * , 046006 ( 2010 ) [ arxiv:1006.1263 [ hep - th ] ] .
s. ryu and t. takayanagi , `` holographic derivation of entanglement entropy from ads / cft , '' phys .
* 96 * , 181602 ( 2006 ) [ hep - th/0603001 ] . + v. e. hubeny , m. rangamani and t. takayanagi , `` a covariant holographic entanglement entropy proposal , '' jhep * 0707 * , 062 ( 2007 ) [ arxiv:0705.0016 [ hep - th ] ] . h. casini , m. huerta and r. c. myers , `` towards a derivation of holographic entanglement entropy , '' jhep * 1105 * , 036 ( 2011 ) [ arxiv:1102.0440 [ hep - th ] ] . a. bhattacharyya , m. sharma and a. sinha , `` on generalized gravitational entropy , squashed cones and holography , '' jhep * 1401 * , 021 ( 2014 ) [ arxiv:1308.5748 [ hep - th ] ] .
+ a. bhattacharyya , a. kaviraj and a. sinha , `` entanglement entropy in higher derivative holography , '' jhep * 1308 * , 012 ( 2013 ) [ arxiv:1305.6694 [ hep - th ] ] .
+ a. bhattacharyya and m. sharma , `` on entanglement entropy functionals in higher derivative gravity theories , '' jhep * 1410 * , 130 ( 2014 ) arxiv:1405.3511 [ hep - th ] ] .
j. de boer , m. kulaxizi and a. parnachev , `` holographic entanglement entropy in lovelock gravities , '' jhep * 1107 * , 109 ( 2011 ) [ arxiv:1101.5781 [ hep - th ] ] .
+ l. y. hung , r. c. myers and m. smolkin , `` on holographic entanglement entropy and higher curvature gravity , '' jhep * 1104 * , 025 ( 2011 ) [ arxiv:1101.5813 [ hep - th ] ] .
s. talaganis , t. biswas and a. mazumdar , `` towards understanding the ultraviolet behavior of quantum loops in infinite - derivative theories of gravity , '' class .
* 32 * , no .
21 , 215017 ( 2015 ) [ arxiv:1412.3467 [ hep - th ] ] . c. imbimbo , a. schwimmer , s. theisen and s. yankielowicz , `` diffeomorphisms and holographic anomalies , '' class .
quant . grav .
* 17 * , 1129 ( 2000 ) [ hep - th/9910267 ] .
+ a. schwimmer and s. theisen , `` entanglement entropy , trace anomalies and holography , '' nucl .
b * 801 * , 1 ( 2008 ) [ arxiv:0802.1017 [ hep - th ] ] .
s. nojiri , s. d. odintsov and s. ogushi , `` cosmological and black hole brane world universes in higher derivative gravity , '' phys .
d * 65 * ( 2002 ) 023521 [ hep - th/0108172 ] .
+ m. cvetic , s. nojiri and s. d. odintsov , `` black hole thermodynamics and negative entropy in de sitter and anti - de sitter einstein - gauss - bonnet gravity , '' nucl .
b * 628 * ( 2002 ) 295 [ hep - th/0112045 ] . | in arxiv:1508.01343 [ hep - th ] , one of the authors proposed that in ads / cft the gravity dual of the boundary @xmath0-theorem is the second law of thermodynamics satisfied by causal horizons in ads and this was verified for einstein gravity in the bulk . in this paper
we verify this for higher derivative theories .
we pick up theories for which an entropy expression satisfying the second law exists and show that the entropy density evaluated on the causal horizon in a rg flow geometry is a holographic c - function .
we also prove that given a theory of gravity described by a local covariant action in the bulk a sufficient condition to ensure holographic c - theorem is that the second law of causal horizon thermodynamics be satisfied by the theory .
this allows us to explicitly construct holographic c - function in a theory where there is curvature coupling between gravity and matter and standard null energy condition can not be defined although second law is known to hold .
based on the duality between c - theorem and the second law of causal horizon thermodynamics proposed in arxiv:1508.01343 [ hep - th ] and the supporting calculations of this paper we conjecture that every unitary higher derivative theory of gravity in ads satisfies the second law of causal horizon thermodynamics .
if this is not true then c - theorem will be violated in a unitary lorentz invariant field theory . |
when dealing with the three - body variational problem with hylleraas basis , it is usually necessary to make extensive use of integrals of the general form @xcite @xmath7 where @xmath8 , @xmath9 and @xmath10 . for the case of @xmath1 , @xmath2 and @xmath3
non - negative ( that is , non - negative powers of @xmath11 , @xmath12 and @xmath13 once the volume element has been taken into account ) , powerful , simple and stable recurrence relations that permit the numerical calculation of these integrals can be found in the literature @xcite .
however , it is sometimes essential to have also an expression for one of the integer indices being negative .
for instance , that happens in the atomic problem when one wants to consider the mean value of the @xmath14 operator @xcite or relativistic corrections @xcite ; or in the nuclear problem when non - local terms are included in a yukawa - like interaction @xcite . in some cases ,
the integrals must be computed in every step of the non - linear optimization procedure , and hence it is clear the need of having a quick and reliable algorithm to compute them .
the specific cases @xmath15 and @xmath16 were already considered in refs .
@xcite and @xcite respectively . for @xmath17
much work has been done @xcite , also including explicitly the coupling of the angular momentum of the two dynamical particles @xcite .
some work has been devoted to the analogous integrals for four- or more - body problems @xcite .
the method proposed in this work to obtain the integrals ( [ i ] ) is specially useful when the same exponential coefficients @xmath4 , @xmath5 and @xmath6 appear in several elements of the variational basis .
to study the general properties of the integral ( [ i ] ) for @xmath1 , @xmath2 and @xmath3 ( possibly negative ) integer numbers and @xmath4 , @xmath5 and @xmath6 real it is convenient to make use of perimetric coordinates @xcite , @xmath19{lcr } u & = & - r_1 + r_2 + r_{12}\\ v & = & r_1 - r_2 + r_{12}\\ w & = & r_1 + r_2 - r_{12}\\ \end{array } \right\ } \label{perimetric}\ ] ] in terms of which the initial integral reads @xmath20 where @xmath21 the integral @xmath22 is explicitly invariant under permutation of conjugated pairs of parameters @xmath23 , and therefore @xmath24 symmetry that will be used throughout this work .
the long range convergence of @xmath22 is ensured if @xmath25 , @xmath26 and @xmath27 are positive real numbers , that is , if @xmath28 that means that one of the exponentials parameters , @xmath4 , @xmath5 or @xmath6 , can be zero or negative , provided that the other two are bigger than the absolute value of the former . note also
that one of the exponential coefficients of @xmath22 can be zero if the power of the corresponding integration variable is negative and high enough .
for instance , @xmath29 with @xmath30 and @xmath31 would yield a convergent result .
anyhow , this is an almost useless case for the variational problem , because for higher power integrals ( that very likely should also be considered ) @xmath29 would lead to divergent quantities . from now on , we assume that the requirements ( [ long_range ] ) are fulfilled
. the study of the short range convergence can be straightforwardly done case by case .
summarizing , for @xmath1 , @xmath2 and @xmath3 integer , and @xmath4 , @xmath5 and @xmath6 real such that the conditions ( [ long_range ] ) are fulfilled , the integral ( [ i ] ) is convergent if and only if @xmath32 to have a procedure to generate the whole set of integrals ( [ i ] ) one needs relations for the cases @xmath33 and @xmath34 where @xmath1 and @xmath2 are non - negative .
as soon as we have checked that the integral we are looking for is convergent , integration over one parameter can be applied to lower the conjugated power , @xmath35 on the other hand , derivation can always be used to increase indices , @xmath36 these properties , together with @xmath37 are useful to derive all the integrals .
note also that for @xmath38 @xmath39 that is , for given @xmath1 , @xmath2 and @xmath3 , @xmath40 is a homogeneous function of @xmath4 , @xmath5 and @xmath6 .
this fact , together with properties ( [ genprosym ] ) and ( [ genpro2 ] ) yields a quite general recursion . indeed , differentiating with respect to @xmath41 in the equation above one gets the recurrence relation @xmath42 valid for well defined integrals , in our case @xmath1 , @xmath2 and @xmath3 fulfilling conditions ( [ genprolmn ] ) . in general
, this recursion is of little utility , for to use it downwards , which is the obvious direction , one would have to know the value of the integrals on a plane @xmath43 .
we will take profit of a particular case of eq .
( [ genprorec ] ) in section [ secmm ] .
for the family of integrals @xmath46 with @xmath47 , a variation of the method exposed in ref.@xcite can be applied . the recurrence relation that one gets is the following , @xmath48\ , , \label{im1rec}\ ] ] where @xmath49 which is a symmetric function under @xmath50 exchange , can be obtained through the relation @xmath51 \ , .
\label{blmrecuns}\ ] ] here the function @xmath52 reads @xmath53 and is defined so that the recursion ( [ blmrecuns ] ) holds also for @xmath54 although @xmath55 and @xmath56 are divergent .
note that @xmath52 is antisymmetric under @xmath57 .
unfortunately , in the recursion ( [ blmrecuns ] ) subtractions are involved , and hence one must look over the stability against roundoff , in particular when @xmath4 and @xmath5 are close to each other .
it is also possible to relate @xmath58 to gauss hypergeometric function , @xmath59 @xcite , yielding @xmath60 where @xmath61 .
the use of the integral representation of the hypergeometric function gives @xmath62 from the definition ( [ blmdef ] ) it is possible to prove the equation @xmath63 plugging this relation in eq .
( [ blmrecuns ] ) yields @xmath64 valid for @xmath65 and @xmath66 .
this equation permits to lower one unit the index @xmath1 of @xmath58 with numerical stability if @xmath67 . in the opposite case
, the symmetry of @xmath58 can be used to lower the index @xmath2 ( see fig .
[ f - stab ] ) . on the other hand , using the gauss relations for contiguous hypergeometric functions one obtains @xmath68 where @xmath69 .
this relation defines a recursion that can be used to move on the diagonals @xmath70 . as it is shown in fig .
[ f - stab ] , the straight line @xmath71 on the @xmath1-@xmath2-plane separates the stability regions of the recursion ( [ blmrecd ] ) , so that one can move with stability from this line in diagonal steps .
the final recipe to compute the set of @xmath72 for @xmath73 is the following ( see fig .
[ f - stab ] ) . first
, two @xmath74 s are to be computed numerically to the required accuracy , namely @xmath75 , 2 n - \left[\case{2 n}{1+\xi}\right ] \right ) \text { and } b\left(\left[\case{2 n}{1+\xi}\right]+1 , 2 n - \left[\case{2 n}{1+\xi}\right ] - 1\right ) \label{2bs}\ ] ] ( respectively , points @xmath76 and @xmath77 in fig .
[ f - stab ] ) .
then the recursion ( [ blmrecd ] ) is used to generate all needed starting points to use the recursion ( [ blmrecl ] ) leftwards ( downwards ) if @xmath67 ( @xmath78 ) .
finally , the @xmath74 s obtained in this way are introduced in eq .
( [ im1rec ] ) . to generate the two initial @xmath74 s one can compute the integral in eq .
( [ blmint ] ) by gauss - legendre quadrature . to optimize the computation of the quadrature a change of variable is needed .
first , we use the symmetry of @xmath58 to render @xmath79 .
next , we apply in eq .
( [ blmint ] ) the change of variable ( @xmath80 ) @xmath81 for values of @xmath82 greater than 0.99 the hypergeometric function can be computed using the transformation formula 15.3.11 of the ref .
@xcite . with the prescription above
, more than fifteen stable figures were obtained using 32 gauss - legendre points for @xmath83 .
note that the prescription ( [ b - chov ] ) has been optimized for the computation of the two initial @xmath74 s given in the expression ( [ 2bs ] ) , and will not provide a similar accuracy for arbitrary values of @xmath1 and @xmath2 .
the particular case @xmath84 is specially simple . indeed , in that case the @xmath74 function to be included in eq .
( [ im1rec ] ) is @xmath85 and the calculations are numerically stable .
the case @xmath84 is not only a mere academic example . in many practical problems the variational basis
is chosen so that any element has the same exponential coefficient both for the coordinates @xmath11 and @xmath12 . if the physical problem requires to deal with @xmath33 integrals , then it is sensible to check whether such a basis can produce the required accuracy .
this selection was successfully used in the context of a nuclear theory problem @xcite . in table [
t - im ] we give some particular values of @xmath46 with fourteen significant figures to provide the reader with checking points .
to generate the set of integrals @xmath34 use can be made of the relation @xmath88 which is valid for @xmath89 .
this equation is easily obtained as a particular case of recurrence ( [ genprorec ] ) . for @xmath30 direct calculation
yields @xmath90 \label{im1m1l0}\ ] ] ( see also ref .
@xcite ) , where @xmath91 is the dilogarithm function . as said , the expression ( [ genprorec ] ) is not applicable for the case @xmath92 . instead
, one gets @xmath93 to fix the constant in the right hand side , we had to make explicit use of the expression ( [ im1m1l0 ] ) .
the recursion ( [ im1m1rec ] ) , which in general is numerically unstable upwards , can be used with stability to decrease the index @xmath1 if @xmath94 , which is the interesting case in physics .
but then , one needs as starting point the integral with the highest wanted @xmath1 . as it can be derived from eqs .
( [ genpro1]-[genpro3 ] ) , that integral can be obtained through the computation of the quadrature @xmath95 where we have defined @xmath96 \right\}\ , .
\nonumber\\ \label{gl}\end{aligned}\ ] ] note that the integrand is positive , and that the sum in the function @xmath97 is very efficiently computed in a single loop . for values of @xmath4 , @xmath5 and @xmath6 of the same order of magnitude
the quadrature converges very quickly for not very small values of @xmath1 ( @xmath98 ) .
this is not the case when one of the parameters is larger than the other , but then a simple change of variable helps to recover convergence .
for instance , the following prescription of changes of variable ( @xmath80 ) @xmath99 produces the accuracies shown in table [ t - acc - g ] . for @xmath100 ( respectively , 20 , 30 and 40 ) more than 3800 integrals per second ( respectively , 2500 , 1800 and 1400 ) were obtained with more than fourteen stable figures ( 32 gauss - legendre points , see table [ t - acc - g ] ) in an inexpensive computer ( a pc with a 200 mhz processor ) .
the prescription ( [ imm - chov ] ) is not useful if both @xmath5 and @xmath6 are much smaller than @xmath4 ( e.g. , @xmath101 ) .
however , the integrals for this case can be safely generated without significant loss of accuracy using the recursion ( [ im1m1rec ] ) upwards .
the relative error in the @xmath1-th integral accumulated because of cancellations , which grows with @xmath1 , can be then approximated as @xmath102 for example , @xmath103 , and the relative error due to cancellations in the repeatedly use of ( [ im1m1rec ] ) to obtain this integral is only of about three times the machine - precision . for smaller values of @xmath5 and @xmath6
the accuracy is bigger .
note that eq .
( [ im1m1l0 ] ) is not appropriate to evaluate @xmath16 when @xmath5 and @xmath6 are smaller than @xmath4 .
a better expression for this case is @xmath104\ , , \label{im1m1l0p}\end{aligned}\ ] ] where the dilogarithm function can be computed through the expansion @xmath105 . for @xmath106 ,
corresponding to @xmath107 , eight terms in this expansion are enough to obtain sixteen stable figures .
a few particular cases of @xmath34 are readily obtained from the recursion ( [ im1m1rec ] ) .
indeed , for @xmath108 and @xmath109 we have @xmath110\,.\ ] ] the specific case @xmath111 reads @xmath112 where the coefficients @xmath113 are to be computed only once . for @xmath114
one does not need more than 52 terms to achieve sixteen stable figures in @xmath115 without using any numerical procedure to accelerate convergence .
the first of these coefficients is @xmath116 .
finally , for @xmath117 one has @xmath118 some values of @xmath119 are presented in table [ t - imm ] .
some recurrence relations to compute the integrals ( [ i ] ) for all negative integer parameters ( @xmath1 , @xmath2 and @xmath3 ) have been presented .
the stability of these recursions has been investigated , and algorithms have been given to use them without loss of accuracy due to cancellations .
the integrals @xmath46 ( where @xmath120 ) can be generated at low computing cost . for the integrals
@xmath119 a quadrature involving @xmath121 terms is needed , where @xmath122 is the highest required @xmath1 and @xmath4 is assumed to be positive . specially simple algorithms are given for the cases @xmath123 , @xmath124 , @xmath125 and @xmath117 .
the author gratefully thanks l.l .
salcedo for helpful comments on a previous version of the manuscript , and e. buenda for some references .
this work was supported by the _ direccin general de enseanza superior _
( spanish education and culture ministry ) through a postdoctoral grant and the project pb95 - 1204 .
9 j. l. calais , and p. o. lwdin , j. mol .
* 8 * , 203 ( 1962 ) .
r. a. sack , c. c. j. roothaan , and w. kolos , j. math .
* 8 * , 1093 ( 1967 ) .
a. j. thakkar , and v.h .
smith , jr .
, phys . rev .
* a15 * , 1 ( 1977 ) . g. breit , phys . rev . *
34 * , 553 ( 1929 ) .
j. caro , c. garca recio , and j. nieves , preprint nucl - th 9801065 , submitted to nuc
. phys . * a*. d. m. fromm , and r. n. hill , phys . rev .
* a36 * , 1013 ( 1987 ) .
p. j. roberts , j. chem .
* 43 * 3547 ( 1965 ) .
n. solony , c. s. lin , and f. w. birss , j. chem . phys .
* 45 * , 976 ( 1966 ) .
l. hambro , phys . rev . *
a5 * , 2027 ( 1972 ) .
g. f. thomas , f. javor , and s. m. rothstein , j. chem . phys . * 64 * , 1574 ( 1976 ) .
f. w. king , phys .
* a44 * , 7108 ( 1991 ) .
yan , and g. w. f. drake , chem .
lett . * 259 * , 96 ( 1996 ) .
i. porras , and f. w. king , phys .
* a49 * , 1637 ( 1994 ) .
yan , and g. w. f. drake , j. phys .
* b30 * , 4723 ( 1997 ) . c. l. pekeris , phys .
rev . * 112 * , 1649 ( 1958 ) .
m. abramowitz , and i. a. stegun , _ handbook of mathematical functions _ ( dover publications , new york , 1972 ) .
c*4c + @xmath126 & @xmath127 & @xmath128 & @xmath129 & @xmath130 + ( 0.05,0.05 ) & 2.5097803893512 ( @xmath13107 ) & 1.1550294761619 ( @xmath13114 ) & 5.8496691184166 ( @xmath13145 ) & 3.7263930880406 ( @xmath13165 ) + ( 0.05,0.20 ) & 1.0069028364788 ( @xmath13107 ) & 4.9856655589152 ( @xmath13112 ) & 1.1263701557234 ( @xmath13141 ) & 1.5245149068463 ( @xmath13164 ) + ( 0.05,1.00 ) & 2.1231968038912 ( @xmath13106 ) & 6.1362890397889 ( @xmath13111 ) & 1.8814741050023 ( @xmath13139 ) & 2.5445500764172 ( @xmath13163 ) + ( 0.05,2.00 ) & 1.0610620020573 ( @xmath13106 ) & 3.0134288353911 ( @xmath13111 ) & 9.0355811977946 ( @xmath13138 ) & 1.2652815721954 ( @xmath13163 ) + ( 0.05,5.00 ) & 4.2435098606331 ( @xmath13105 ) & 1.1992798691243 ( @xmath13111 ) & 3.5754444925655 ( @xmath13138 ) & 5.0533341605295 ( @xmath13162 ) + ( 0.20,0.05 ) & 5.9511548712149 ( @xmath13106 ) & 1.4149826065364 ( @xmath13112 ) & 7.4333291536966 ( @xmath13139 ) & 4.6331742767202 ( @xmath13161 ) + ( 0.20,0.20 ) & 2.5245340864963 ( @xmath13106 ) & 3.8096117209107 ( @xmath13111 ) & 1.5683171148516 ( @xmath13138 ) & 1.3617534581186 ( @xmath13161 ) + ( 0.20,1.00 ) & 4.9479973346792 ( @xmath13105 ) & 7.1119098709184 ( @xmath13110 ) & 1.5019778106939 ( @xmath13137 ) & 2.7821828122703 ( @xmath13160 ) + ( 0.20,2.00 ) & 2.4501414762596 ( @xmath13105 ) & 3.5414728355891 ( @xmath13110 ) & 7.3432360158513 ( @xmath13136 ) & 1.3921604053583 ( @xmath13160 ) + ( 0.20,5.00 ) & 9.7730721646852 ( @xmath13104 ) & 1.4149098737738 ( @xmath13110 ) & 2.9191066335088 ( @xmath13136 ) & 5.5698482446321 ( @xmath13159 ) + ( 1.00,0.05 ) & 2.3506936424270 ( @xmath13105 ) & 2.1738100662595 ( @xmath13108 ) & 1.8638785541149 ( @xmath13130 ) & 1.3896745377911 ( @xmath13153 ) + ( 1.00,0.20 ) & 6.8960382056483 ( @xmath13104 ) & 9.7511110023298 ( @xmath13107 ) & 7.0043082969132 ( @xmath13129 ) & 2.2092326515384 ( @xmath13151 ) + ( 1.00,1.00 ) & 2.6754225852273 ( @xmath13103 ) & 2.0193624551498 ( @xmath13107 ) & 1.4992836690734 ( @xmath13129 ) & 1.5383073655665 ( @xmath13149 ) + ( 1.00,2.00 ) & 9.7180274971942 ( @xmath13102 ) & 1.0013416984929 ( @xmath13107 ) & 7.5130229486046 ( @xmath13128 ) & 6.9573280539039 ( @xmath13148 ) + ( 1.00,5.00 ) & 3.5921654332679 ( @xmath13102 ) & 3.9932069716804 ( @xmath13106 ) & 3.0070041814429 ( @xmath13128 ) & 2.7138972320215 ( @xmath13148 ) + ( 2.00,0.05 ) & 5.6374109387576 ( @xmath13104 ) & 1.4250979709052 ( @xmath13106 ) & 9.1546004403515 ( @xmath13124 ) & 2.9503794635963 ( @xmath13149 ) + ( 2.00,0.20 ) & 1.5196492752737 ( @xmath13104 ) & 4.7607751878871 ( @xmath13105 ) & 1.4095247283656 ( @xmath13124 ) & 1.9130948733180 ( @xmath13147 ) + ( 2.00,1.00 ) & 1.5320840275499 ( @xmath13102 ) & 3.8998899902794 ( @xmath13104 ) & 7.9782256684551 ( @xmath13122 ) & 8.8675800592916 ( @xmath13140 ) + ( 2.00,2.00 ) & 1.6829755898961 ( @xmath13101 ) & 1.6286558844196 ( @xmath13104 ) & 3.5727727922307 ( @xmath13122 ) & 8.6310437705915 ( @xmath13139 ) + ( 2.00,5.00 ) & 4.3389324770986 ( @xmath13100 ) & 6.1452543996272 ( @xmath13103 ) & 1.3850914702513 ( @xmath13122 ) & 2.9003763096969 ( @xmath13139 ) + ( 5.00,0.05 ) & 8.9272368511676 ( @xmath13103 ) & 2.0896049079214 ( @xmath13103 ) & 1.2958485756621 ( @xmath13118 ) & 4.7703597474580 ( @xmath13144 ) + ( 5.00,0.20 ) & 2.3569834307325 ( @xmath13103 ) & 5.6914967826940 ( @xmath13102 ) & 1.0121437234633 ( @xmath13117 ) & 2.4003665222216 ( @xmath13142 ) + ( 5.00,1.00 ) & 1.5138128168377 ( @xmath13101 ) & 6.1945864757004 ( @xmath13100 ) & 2.8220667212483 ( @xmath13113 ) & 8.7853148834688 ( @xmath13133 ) + ( 5.00,2.00 ) & 3.2220328921320 ( @xmath13201 ) & 5.5803922147916 ( @xmath13201 ) & 1.0214473659965 ( @xmath13112 ) & 1.1612978333117 ( @xmath13128 ) + ( 5.00,5.00 ) & 3.4970794973642 ( @xmath13203 ) & 1.0465458754289 ( @xmath13201 ) & 2.1723867163698 ( @xmath13111 ) & 1.6228394644288 ( @xmath13124 ) + d*4d @xmath1 & @xmath133=10 & @xmath133=16 & @xmath133=24 & @xmath133=32 + 5 & 5.9 & 8.9 & 11.1 & 12.6 + 6 & 6.1 & 9.4 & 12.1 & 13.9 + 7 & 5.6 & 9.6 & 12.9 & 14.9 + 8 & 5.6 & 9.7 & 13.6 & 15.4 + 9 & 5.9 & 10.3 & 14.1 & 15.3 + 10 & 5.4 & 9.3 & 14.4 & 15.3 + 15 & 5.1 & 8.7 & 14.0 & 15.2 + 20 & 5.2 & 8.3 & 13.6 & 15.0 + 25 & 4.7 & 8.2 & 12.9 & 14.9 + 30 & 5.0 & 8.1 & 12.1 & 14.8 + 35 & 4.7 & 7.5 & 11.4 & 14.8 + 40 & 4.5 & 7.0 & 11.0 & 14.6 + c*4c + @xmath126 & @xmath100 & @xmath134 & @xmath135 & @xmath136 + ( 0.01,0.01 ) & 6.6191662489867 ( @xmath13106 ) & 3.6251447231256 ( @xmath13118 ) & 3.2983926561168 ( @xmath13132 ) & 8.5733728363337 ( @xmath13147 ) + ( 0.01,0.20 ) & 1.7540859652228 ( @xmath13106 ) & 5.3112685642066 ( @xmath13117 ) & 3.3942282862188 ( @xmath13131 ) & 6.9857465097091 ( @xmath13146 ) + ( 0.01,0.50 ) & 6.8609259644037 ( @xmath13105 ) & 2.0155302235322 ( @xmath13117 ) & 1.3180581225243 ( @xmath13131 ) & 2.7475204208958 ( @xmath13146 ) + ( 0.01,1.00 ) & 3.3243271983196 ( @xmath13105 ) & 9.9937617391729 ( @xmath13116 ) & 6.5672190824270 ( @xmath13130 ) & 1.3709460938605 ( @xmath13146 ) + ( 0.01,2.00 ) & 1.6469713752668 ( @xmath13105 ) & 4.9876597990304 ( @xmath13116 ) & 3.2808597542740 ( @xmath13130 ) & 6.8512909302258 ( @xmath13145 ) + ( 0.01,5.00 ) & 6.5729176667975 ( @xmath13104 ) & 1.9940611043105 ( @xmath13116 ) & 1.3120390818053 ( @xmath13130 ) & 2.7401331921302 ( @xmath13145 ) + ( 0.01,10.00 ) & 3.2854418466598 ( @xmath13104 ) & 9.9695938948872 ( @xmath13115 ) & 6.5599782224617 ( @xmath13129 ) & 1.3700392619726 ( @xmath13145 ) + ( 0.20,0.20 ) & 4.5604766596349 ( @xmath13105 ) & 3.5127689695513 ( @xmath13116 ) & 4.9063595251091 ( @xmath13129 ) & 2.0412224928657 ( @xmath13144 ) + ( 0.20,0.50 ) & 1.4318580613654 ( @xmath13105 ) & 7.2132201828707 ( @xmath13115 ) & 8.1925006515660 ( @xmath13128 ) & 3.0132668197157 ( @xmath13143 ) + ( 0.20,1.00 ) & 6.1814890074087 ( @xmath13104 ) & 3.2566277622792 ( @xmath13115 ) & 3.8027988040087 ( @xmath13128 ) & 1.4138666122607 ( @xmath13143 ) + ( 0.20,2.00 ) & 2.9651034471161 ( @xmath13104 ) & 1.5963963424310 ( @xmath13115 ) & 1.8718211282572 ( @xmath13128 ) & 6.9707881275722 ( @xmath13142 ) + ( 0.20,5.00 ) & 1.1742458080974 ( @xmath13104 ) & 6.3522447488220 ( @xmath13114 ) & 7.4555535110623 ( @xmath13127 ) & 2.7776640095065 ( @xmath13142 ) + ( 0.20,10.00 ) & 5.8633250910678 ( @xmath13103 ) & 3.1737770991220 ( @xmath13114 ) & 3.7255334615001 ( @xmath13127 ) & 1.3880777482000 ( @xmath13142 ) + ( 0.50,0.50 ) & 2.7602780782403 ( @xmath13104 ) & 2.1429404549106 ( @xmath13114 ) & 3.1095312824899 ( @xmath13126 ) & 1.3602027847527 ( @xmath13140 ) + ( 0.50,1.00 ) & 7.8596287095799 ( @xmath13103 ) & 4.2261129342848 ( @xmath13113 ) & 5.2077776493094 ( @xmath13125 ) & 2.0657542275279 ( @xmath13139 ) + ( 0.50,2.00 ) & 3.2906652367190 ( @xmath13103 ) & 1.8845025819044 ( @xmath13113 ) & 2.3675994785514 ( @xmath13125 ) & 9.4583791413986 ( @xmath13138 ) + ( 0.50,5.00 ) & 1.2667740672550 ( @xmath13103 ) & 7.3501151392020 ( @xmath13112 ) & 9.2602020503771 ( @xmath13124 ) & 3.7039195719454 ( @xmath13138 ) + ( 0.50,10.00 ) & 6.3029915005903 ( @xmath13102 ) & 3.6623979936588 ( @xmath13112 ) & 4.6157947923024 ( @xmath13124 ) & 1.8465337659004 ( @xmath13138 ) + ( 1.00,1.00 ) & 9.2233071085062 ( @xmath13102 ) & 3.9012855025163 ( @xmath13111 ) & 3.1367380768282 ( @xmath13122 ) & 7.6500307975383 ( @xmath13134 ) + ( 1.00,2.00 ) & 2.1119403628357 ( @xmath13102 ) & 6.5569071489222 ( @xmath13110 ) & 4.6001747109760 ( @xmath13121 ) & 1.0314129070250 ( @xmath13134 ) + ( 1.00,5.00 ) & 7.2422587486309 ( @xmath13101 ) & 2.3596051913831 ( @xmath13110 ) & 1.6728442527147 ( @xmath13121 ) & 3.7666888585022 ( @xmath13133 ) + ( 1.00,10.00 ) & 3.5623476884278 ( @xmath13101 ) & 1.1649070967644 ( @xmath13110 ) & 8.2662607382138 ( @xmath13120 ) & 1.8620745501178 ( @xmath13133 ) + ( 2.00,2.00 ) & 8.9219060942223 ( @xmath13100 ) & 6.4071580456640 ( @xmath13107 ) & 8.8443126974508 ( @xmath13116 ) & 3.7177770572299 ( @xmath13127 ) + ( 2.00,5.00 ) & 1.3397156572731 ( @xmath13100 ) & 7.4723481641427 ( @xmath13106 ) & 9.1598047862917 ( @xmath13115 ) & 3.5719240038509 ( @xmath13126 ) + ( 2.00,10.00 ) & 6.2645017020798 ( @xmath13201 ) & 3.5448536509329 ( @xmath13106 ) & 4.3596533454989 ( @xmath13115 ) & 1.7025891413660 ( @xmath13126 ) + ( 5.00,5.00 ) & 3.7697615446040 ( @xmath13203 ) & 2.6042296175688 ( @xmath13101 ) & 3.4861019106182 ( @xmath13107 ) & 1.4250284687126 ( @xmath13115 ) + ( 5.00,10.00 ) & 6.8653568463440 ( @xmath13204 ) & 3.7115610730907 ( @xmath13100 ) & 4.4359819939145 ( @xmath13106 ) & 1.6880938843748 ( @xmath13114 ) + ( 10.00,10.00 ) & 4.5242997355095 ( @xmath13206 ) & 7.2459128782254 ( @xmath13205 ) & 2.2554462512897 ( @xmath13201 ) & 2.1460242640172 ( @xmath13104 ) + | stable recurrence relations are presented for the numerical computation of the calais - lwdin integrals @xmath0 ( @xmath1 , @xmath2 and @xmath3 integer , @xmath4 , @xmath5 and @xmath6 real ) when the indices @xmath1 , @xmath2 or @xmath3 are negative .
useful formulas are given for particular values of the parameters @xmath4 , @xmath5 and @xmath6 . |
mt . pleasant observatory is devoting resources to observe glitches at high resolution .
two telescopes with specifically designed equipment are automated for continuous monitoring and glitch reporting of multiple pulsars .
the 14-metre radio telescope monitors the vela pulsar for over 18 hours each day , and when set , psr b1641 - 45 .
both pulsars are known to be young , active and to glitch .
the vela pulsar glitches approximately every 1000 days .
the 14-metre monitors this pulsar for the specific purpose of timing during and around glitch events .
three uncooled receivers are mounted at the prime focus of the 14-metre to allow continuous dispersion measure determination .
two are stacked disk , dual polarisation with central frequencies of 635 mhz and 990 mhz additional to a right hand circular helix at 1391 mhz .
bandwidths are 250 khz , 800 khz and 2 mhz respectively , limiting pulse broadening from interstellar dispersion to less than 1% .
the output is folded for two minutes giving an integrated pulse profile of 1344 pulses .
the backend to the 990 mhz receiver also has incoherent dedispersion across 8 adjacent channels allowing a study of individual pulses .
a new addition is the daily monitoring of 10 other young , active pulsars with the 26-metre radio telescope .
a cryogenic dual polarisation 20 cm receiver with system temperature of 60 k centred at 1522 mhz is fed to a backend recording total intensity from a filterbank of 24 , 2.5 mhz bandwidth channels .
the single pulse data from the latest vela glitch showed a previously unobservable fast decay term ( dodson et al .
the post - spin - up short - term decay is below the resolution of previously reported two minute folds .
using such data would derive a glitch epoch that is too early .
these high time resolution studies show a fast decay that can be modelled by a core response ( figure 1 ) .
it will provide constraints on the nature of the coupling of the stellar crust to the liquid interior ( lewis , link , & dodson in preparation ) .
the current data shows a need to improve the signal - to - noise ratio of the system in order to examine this phenomenon in greater detail .
the 14-metre time of arrival data archive has been analysed with improved telescope position , pulsar position and recent revised proper motion ( legge 2002 ) .
a full list of glitch separation and size has been produced with a wealth of information on minor glitch events to be compared with similar work ( cordes , downs , & krause - polstorff 1988 ) .
the post - glitch behaviour shows there is no simple pattern in the major observed glitches 5 to 13 in vela ( figure 2 ) .
the preliminary result confirms a low braking index ( lyne et al .
1996 ) , but is extremely sensitive to where after the glitches the second derivative of frequency is fitted .
vela is important to discussions of true to characteristic age of pulsars ( gaensler & frail 2000 ; gaensler et al .
2000 ; kaspi et al .
the unique devotion of the 14-metre at mt . pleasant will be taken advantage of in the near future by completing a sampling and recording system to coherently dedisperse a 32 mhz ( both polarisations ) instantaneous bandpass on the 635 mhz centred system .
the coherent dedispersion system is an adaption of the metshovi disk - based vlbi recorder ( ritakari & mujunen 2002 ) with sufficient disk space for two hours of archived data .
the data will be stored on the looping disk for post glitch analysis along side the 990 mhz system that at present automatically alerts us to a glitch event .
the new backend will increase the detected bandwidth at 635 mhz by a factor of 150 .
the decrease in the frontend system temperature , will improve the signal - to - noise by at least an order of magnitude over the current 990 mhz single pulse system .
the receivers , amplifiers , filters , power supplies , backend and software of the 14-metre have been refined and reinstalled , improving the capabilities of the telescope .
improvements of increasing bandwidth , observing time , reliability , number of pulsars monitored , a second larger telescope and reduced system noise are all aimed at observing more glitches , more effectively .
the motivation is to improve timing resolution to gain more information of the short time - scale spin - up events similar to that observed in 2000 . an important consequence of the upgrades to the mt .
pleasant automated system is the turn - around time for glitch reporting is now hours ( day and night ) . during the 2000 event
, this allowed a target - of - opportunity observation on the chandra x - ray observatory ( helfand et al .
future events will be in conjunction with more formal too requests to find heating of the stellar crust .
glitches must be observed with higher timing resolution to study the short - term decay in more detail .
this provides an insight to the interior physics of neutron stars and equations of state .
as glitches are not precisely predictable , a continuous monitoring program is required .
the new and improved programs in hobart increase the chances of observing glitches in real - time .
continuous monitoring when combined with glitch alerting allows observation of the post - glitch heating of the crust .
post - glitch heating is believed to be resultant from the energy deposition during a glitch ( vanriper , epstein , & guy 1991 ; hirano et al .
1997 ) .
x - ray observations of the neutron star as close as possible to the glitch epoch are required to show all heat dissipation time - scales . a study of the amount of heat energy dissipated over time will allow further insights into neutron star structure ( hirano et al . 1997 ; vanriper et al .
1991 ) .
cordes j. m. , downs g. s. , & krause - polstorff j. 1988 , , 330 , 847 dodson r. g. , mcculloch p. m. , & lewis d. r. 2002 , , 564 , l85 gaensler b. m. , frail d. a. ; 2000 , nature , 406 , 158 gaensler b. m. , stappers b. w. , frail d. a. , moffett d. a. , johnston s. , chatterjee s. 2000 , , 318 , 58 helfand d. j. , gotthelf e. v. , & halpern j. p. 2001
, , 556 , 380 hirano s. , shibazaki n. , umeda h. , & nomoto k. 1997 , , 491 , 286 kaspi v. m. , roberts m. e. , vasisht g. , gotthelf e. v. , pivovaroff m. , & kawai n. 2001 , , 560 legge d. r. 2002 , phd thesis , university of tasmania lyne a. g. , pritchard r. s. , graham - smith f. , & camilo f. 1996 , nature , 381 , 497 ritakari j. , mujunen a. 2002 , international vlbi service for geodesy and astrometry , general meeting proceedings tsukuba , japan , 2002 | two daily pulsar monitoring programs are progressing at the mount pleasant observatory , hobart , tasmania , australia . a new system involving the 26-metre radio telescope monitors 10 young pulsars daily and is focussed on near - real - time glitch finding .
this will allow target of opportunity observations to measure post - glitch heating of the neutron star surface ( helfand , gotthelf , & halpern 2001 ) .
the 14-metre continues its 21st year of daily monitoring of the vela pulsar with a recent comprehensive frontend upgrade .
this is prior to an upgrade of the backend equipment currently in progress .
the 14-metre observed the most recent glitch of the vela pulsar in january 2000 to the highest time resolution of any glitch and revealed a particularly short - term decay component ( dodson , mcculloch , & lewis 2002 ) .
this decay component will provide constraints to the nature of the coupling of the stellar crust to the liquid interior .
# 1_#1 _ # 1_#1 _ = # 1 1.25 in .125 in .25 in |
since the discovery of the colossal magnetoresistance ( cmr ) effect , mn oxides with a perovskite structure and their derivatives have been intensively investigated . in particular ,
two - dimensional ferromagnetic ( fm ) metal ( la , sr)@xmath2mn@xmath13o@xmath14 has been attracting much attention because such metals are rarely reported .
sr@xmath2fe@xmath3co@xmath4o@xmath5 can be expected as one of such rare materials[1 , 2 ] .
the end material sr@xmath2fe@xmath13o@xmath5 is an antiferromagnet below @xmath15 = 120 k , and changes from a semiconductor to a metal at 350 k[3 ] . with increasing co - concentration
, this system approaches an fm metal and negative magnetoresistance ( mr ) is observed .
ghosh @xmath16 @xmath17
. suggest that the ferromagnetism is driven by spin - dependent electron transfer processes between fe@xmath18 and co@xmath18[4 ] .
since highly oxidized states of both fe@xmath18 and co@xmath18 are unstable , the crystal structure of sr@xmath2fe@xmath3co@xmath4o@xmath5 has a large amount of oxygen vacancies .
the oxygen deficiency @xmath19 of sr@xmath2fe@xmath20co@xmath21o@xmath5 treated under high oxygen pressure of 600 bar at 1173 k is 0.20 @xmath22 0.01[4 ] .
a decrease in the average valence of @xmath23-site cations weakens the spin - dependent electron transfer process , which probably leads to coexistence of fm and afm ( or pm ) phases .
such phase coexistence can be expected to enhance the mr of sr@xmath2fe@xmath3co@xmath4o@xmath5 , as is often the case with cmr manganites . in this study , in order to enhance the mr , we have prepared @xmath1-substituted sr@xmath2fe@xmath11co@xmath12o@xmath5 and have investigated the physical properties .
sr@xmath24fe@xmath11co@xmath12o@xmath5 was prepared in a polycrystalline form by a solid - state reaction method .
mixed powders of srco@xmath2 , @xmath25o@xmath2 ( @xmath1 = la , nd , and gd ) , fe@xmath2o@xmath26 , and coo were heated at 1173 k in air and sintered at 1673 k in o@xmath13 .
the sample was then annealed under high oxygen pressure of 240 bar at 773 k for 10 h , and cooled down to about 300 k at a rate of 40 k / h .
crystallographic analysis of the obtained samples was performed by x - ray diffraction method at room temperature .
the oxygen content of the synthesized samples was estimated using iodometric titration with an accuracy of @xmath22 0.02 .
the magnetic properties were measured using a quantum design , physical property measurement system ( ppms - 9 t ) .
the resistivity was measured using a standard four - probe method .
figures 1 ( a ) and ( b ) show the temperature dependence of the magnetization ( @xmath27-@xmath28 ) and magnetization curves ( @xmath27-@xmath29 ) of sr@xmath30fe@xmath11co@xmath12o@xmath5 ( @xmath1 = la and gd ) .
the @xmath27-@xmath28 of sr@xmath2fe@xmath11co@xmath12o@xmath5 shows an increase below 250 k. the zero field cooling ( zfc ) process deviates from field cooling ( fc ) below 80 k. the magnetization is fm below 250 k , and the magnetization value reaches about 2.8 @xmath31/@xmath23-site at 50 k , as seen from the @xmath27-@xmath29 curve .
these magnetic behaviors result from the emergence of the cluster glass state , as previously reported[1 ] .
this cluster glass - like behavior is prominent below 80 k. in sr@xmath9la@xmath10fe@xmath11co@xmath12o@xmath5 , the onset temperature of the fm transition shifts to lower temperatures , and the fm moment is slightly suppressed ( fig .
1(b ) ) . in the case of gd@xmath32-substitution , the fm moment is close to that of sr@xmath9la@xmath10fe@xmath11co@xmath12o@xmath5 , but the fm transition temperature is further lower to around 150 k. figures 1 ( c ) and ( d ) show the temperature dependence of the resistivity and mr of sr@xmath30fe@xmath11co@xmath12o@xmath5 . here
the mr value is defined as mr = ( @xmath33(80 koe ) @xmath34 @xmath33(0 koe ) ) / @xmath33(0 koe ) .
the resistivity of sr@xmath2fe@xmath11co@xmath12o@xmath5 is semiconducting , and the negative mr is observed below the onset temperature of ferromagnetism .
the absolute value of mr , @xmath35mr@xmath36 , increases gradually up to @xmath37 % at 5 k with decreasing temperature .
la@xmath32-substitution makes this system insulating , while the mr of @xmath1 = la is slightly enhanced compared with that of sr@xmath2fe@xmath11co@xmath12o@xmath5 at low temperatures . with decreasing the ionic radius of @xmath0 , the resistivity is increasing and the @xmath35mr@xmath36 is further enhanced from @xmath38 % ( nondoped ) to @xmath39 % ( @xmath1 = gd ) .
figures 2 ( a ) and ( b ) show the @xmath27-@xmath29 curve and isothermal mr of sr@xmath9la@xmath10fe@xmath11co@xmath12o@xmath5 at 5 k. a significant difference between la@xmath32-doped and nondoped samples is that a metamagnetic transition is found in the la@xmath32-doped sample around 16 koe .
the metamagnetic transition accompanies an abrupt decrease in the resistivity and is reversibly observed by applying magnetic fields .
similar metamagnetic behavior is observed in all the @xmath0-substituted samples , but not in the nondoped sample .
we discuss the origin of the enhancement of the mr of sr@xmath30fe@xmath11co@xmath12o@xmath5 .
we confirmed that the @xmath19 of the @xmath0-substituted sample almost coincides with that of sr@xmath2fe@xmath11co@xmath12o@xmath5 by iodometric titration .
in other words , @xmath0-substitution merely decreases the average valence of the @xmath23-site cation .
an oxygen - vacancy ordered structure ( or oxygen stoichiometry ) might exist in the vicinity of @xmath19 = 0.20 .
such a decrease in the average valence of the @xmath23-site cation probably creates antiferromagnetic ( afm ) ( or non - fm ) clusters around @xmath0 .
it is therefore reasonable to conclude that the suppression of ferromagnetism by @xmath0-substitution is due to the creation of these afm ( or non - fm ) clusters .
the volume reduction of the fm clusters and randomness due to @xmath0-substitution cause the resistivity to be insulating .
the fm clusters are aligned parallel to applied magnetic fields . as a result , the large negative mr is observed .
in addition , metamagnetism accompanying a reduction in the resistivity also contributes to the enhancement of the mr at low temperatures .
we have investigated a quasi two - dimensional ferromagnet sr@xmath2fe@xmath11co@xmath12o@xmath5 partially substituted by @xmath0 .
this substitution of @xmath0 suppresses ferromagnetism and causes the system to be more insulating , which are attributed to an increase in afm clusters created around @xmath0 . in sr@xmath9gd@xmath10fe@xmath11co@xmath12o@xmath5 ,
the mr value is enhanced by about 20 % compared with that of sr@xmath2fe@xmath11co@xmath12o@xmath5 .
competition between fm and afm ( or non - fm ) clusters is essential for the enhancement of the mr .
one of the plausible explanations for the origin of this enhanced mr is as follows . when applying magnetic fields to sr@xmath9gd@xmath10fe@xmath11co@xmath12o@xmath5 , the fm clusters are aligned in the same direction , resulting in a reduction in resistivity .
in addition , in the @xmath0-doped samples , a metamagnetic transition accompanying an abrupt decrease in the resistivity occurs at low temperatures , which also contributes to the enhancement of the mr of sr@xmath30fe@xmath11co@xmath12o@xmath5 .
this work was partly supported by the asahi glass foundation , and grant - in - aid for scientific research ( c ) from the japan society for promotion of science . | we have investigated the @xmath0-substitution ( @xmath1 = lanthanoid ) effect of a quasi two - dimensional ferromagnet sr@xmath2fe@xmath3co@xmath4o@xmath5 ( @xmath6@xmath7@xmath8 ) . with increasing @xmath0-concentration ,
the ferromagnetism is gradually suppressed and the resistivity is increasing , which are ascribed to an increase in antiferromagnetic ( afm ) clusters created by @xmath0-substitution . in sr@xmath9gd@xmath10fe@xmath11co@xmath12o@xmath5 , the magnetoresistance ( mr ) is enhanced by about 20 % compared with that of sr@xmath2fe@xmath11co@xmath12o@xmath5 .
coexistence of ferromagnetic ( fm ) and afm phases is essential for the enhancement of the mr . applied magnetic fields
align the fm clusters in the same direction , resulting in a reduction in the resistivity .
a metamagnetic transition observed in the @xmath0-doped samples also contributes to the enhancement of the mr . |
in recent years , the results from supernova ia have shown that the expansion of our universe is accelerating unlike the friedmann - robertson - walker ( frw ) cosmological models with nonrelativistic matter and radiation @xcite . also , the cosmic background radiation data imply that our universe is in positively accelerated state @xcite .
the cosmological constant @xmath0 as the usual vacuum energy may be responsible for this accelerating evolution of the universe with the negative pressure .
unfortunately , the measured value of the cosmological constant is 120 orders of magnitude smaller than its theoretical predicted value @xcite . on the other hand , thr chaplygin gas , as a perfect fluid which behaves like a pressureless fluid at early times and a cosmological constant at later times ,
can be a candidate for the dark energy @xcite .
the generalized chaplygin gas with negative pressure is described by an exotic equation of state , @xmath1 where @xmath2 is a positive parameter , @xmath3 is the pressure , @xmath4 is the energy density , and @xmath5 is a positive parameter so that @xmath6 . in the standard model of chaplygin gas , we set @xmath7 @xcite .
recently , various studies have been done in the literature , such as cosmology with the chaplygin gas to explain the transition from a dust - dominated universe to the accelerating expansion stage @xcite , modified generalized chaplygin gas @xcite , modified cosmic chaplygin gas @xcite , and phenomenological relations between brane - world scenarios and frw minisuperspace cosmologies in the presence of the generalized chaplygin gas @xcite .
the idea of an early chaplygin gas phase in the universe was first suggested in refs .
@xcite and later extended in refs .
it is interesting to note that the generalized chaplygin gas has a connection with string theory .
it can be effectively obtained in the light cone parametrization from the nambu - goto action for a @xmath8-brane in a ( @xmath9,1)-dimensional spacetime @xcite .
this motivates the application of the generalized chaplygin gas model at early time in the evolution of the universe .
in particular , the chaplygin inflation is addressed in ref .
@xcite and the tachyon - chaplygin inflationary universe is studied in ref .
@xcite .
in fact , chaplygin spacetime models are related to a generalized born - infeld action with a complex scalar field which corresponds to a perturbed @xmath8-brane in a ( @xmath10,1)-dimensional spacetime @xcite .
a generalized born - infeld phantom inspired generalized chaplygin gas model is presented in ref .
this opens a window to investigate brane - world physics based on a phenomenological point of view .
brane - world scenarios could explain various cosmological effects such as the inflation based on the wmap data @xcite , the origin of our universe , and other fundamental issues such as the hierarchy problem @xcite .
the quantum cosmological creation of brane - world models is also addressed in refs .
this allows us to use the similarities between the quantum cosmology of a brane - world model and the chaplygin gas quantum cosmology .
recently , a single - field inflation scenario is presented in ref .
@xcite where the properties of the inflaton field areequivalent to a generalized chaplygin gas .
based on the measurements released by the planck data and the wmap large - angle polarization , it is found that the @xmath5 parameter is given by @xmath11 @xcite . in ref .
@xcite , a slow - roll inflationary scenario is modeled by a generalized chaplygin gas that can interpolate between a network of frustrated topological defects and a de sitter - like or a power - law inflationary era .
after the universe reached the dust - dominated stage , the exponentially expanding evolution of the universe represents a quantum mechanical transition with some remnant component of the original wave function of the universe @xcite .
in fact , there is a quantum mechanical background behind the classically observed universe on large scales during the dust dominance epoch which is based on the cosmological influence of a rapidly oscillating wave function with a small amplitude .
this wave function remnant can be considered as a robust component with respect to decoherence processes in the early times @xcite . in 1970
, schutz introduced a velocity potential representation for the four - velocity of a perfect fluid in general relativity @xcite .
the equations of hydrodynamics for a perfect fluid are expressed in terms of six scalar potentials @xmath12 , @xmath13 , @xmath14 , @xmath15 , @xmath16 , and @xmath17 so that @xmath18 where each of these potentials has their own equation of motion . indeed , these equations are equivalent to those based on divergence of the stress - energy tensor @xcite . here
, @xmath12 is the specific enthalpy and @xmath17 is the specific entropy of the fluid .
the potentials @xmath14 and @xmath15 are related with rotations and , hence , they are absent in the frw universe due to the symmetry of the model .
the potentials @xmath13 and @xmath16 do not have clear physical meaning .
moreover , the four - velocity obeys the usual normalization , namely , @xmath19 . the velocity potential version of the perfect fluid is based on the variational principle with the following lagrangian density @xmath20 where @xmath21 is the ricci curvature scalar in four dimensions , @xmath3 is the pressure of the perfect fluid , and @xmath22 is determinate of the four - metric .
variation of this action with respect to the metric gives rise to the einstein field equations , and the variation with respect to each of the velocity potentials yields the equations of motion for the potentials @xcite .
also , in this framework , the arnowitt - deser - misner ( adm ) formalism can be used to decompose the spacetime metric @xmath23 in terms of the three - dimensional metric @xmath24 , shift vector @xmath25 , and the lapse function @xmath26 to obtain the hamiltonian and einstein equations in three dimensions @xcite .
a new theory of gravity presented by hoava is based on the asymmetry scaling of space @xmath27 and time @xmath28 , where it is characterized by a scaling parameter @xmath29 and the dynamical critical exponent @xmath30 as @xcite @xmath31 the resulting theory , the so - called hoava - lifshitz ( hl ) gravity , is proved to be power - countable renormalizable .
this theory is based on the assumption that higher spatial - derivative correction terms such as different powers of the spatial curvature and its derivatives could be added to the standard einstein - hilbert action .
this result leads to improvement in the uv behavior of the graviton propagator , but the lorentz invariance as a fundamental symmetry of theory is broken @xcite .
hoava made an important assumption about the lapse function which simplified the hl gravitational action , the so - called `` projectability condition , '' as @xmath32 .
projectable theories lead to a unique integrated hamiltonian constraint , which also cause great complications when they are compared with general relativity . on the other hand , as in general relativity ,
nonprojectable theories give rise to a local hamiltonian constraint @xcite .
however , since frw spacetime is homogeneous and isotropic , the spatial integral can be dropped from the integrated hamiltonian constraint @xcite which results in a true local constraint even for the projectable case .
thus , in our study it is sufficient to consider the frw - hl projectable theory as the starting point .
another assumption that is introduced by hoava is the principle of detailed balance .
based on this condition , the potential in the gravitation action is originated from the gradient flow generated by a three - dimensional action and reduces the number of independent coupling constants .
notice that it has been recently found that the detailed balance condition can also be relaxed @xcite . when the spacetime is asymmetric ( [ eq4 ] ) , the dimensions of space and time are different , namely , @xmath33=[\kappa]^{-1},\qquad[t]=[\kappa]^{-z},\ ] ] where @xmath34 is a placeholder symbol with the dimensions of momentum . for general values of @xmath30 ,
the classical scaling dimensions of the fields are given by @xmath35=[n]=1,\qquad[n^i]=[\kappa]^{z-1},\qquad[ds]=[\kappa]^{-1},\end{aligned}\ ] ] and therefore @xmath36=[\kappa]^{-z - d}$ ] @xcite . throughout the paper , we take @xmath37 .
the gravitational action of the hl model consists of the kinetic part @xmath38 and the potential part @xmath39 in the form @xmath40 .
the kinetic part comes from the einstein - hilbert action and in terms of the adm variables takes the form @xmath41 where @xmath42 is coupling constant and @xmath43 is the exterior curvature tensor , @xmath44 the potential part of the gravitational action is given by @xmath45,\ ] ] where @xmath46 $ ] is a scalar function which depends on the spatial metric and its spatial derivatives @xmath47&=&g_0\zeta^6+g_1\zeta^4r+g_2\zeta^2r^2+g_3\zeta^2r_{ij}r^{ij}\nonumber\\ & & + g_4r^3+g_5rr_{ij}r^{ij}+g_6r^i_{\,\,j}r^j_{\,\,k}r^k_{\,\,i}\nonumber\\\label{eq8 } & & + g_7r\nabla^2r+g_8\nabla_ir_{jk}\nabla^ir^{jk}.\hspace{1cm}\end{aligned}\ ] ] here we used @xmath48 to ensure the coupling constants @xmath49s to be all dimensionless .
the full hl action that we consider throughout the paper is @xmath50 where we set @xmath51 and @xmath52 @xcite .
the solutions of the einstein field equations can be classified with respect to the type of singularities as follows : ( a ) quasiregular singularities : the observer measures no divergent physical quantity , except when its world line arrives at the singularity , e.g. , the conical singularity of a cosmic string .
( b ) scalar curvature singularities : the observer feels diverging tidal forces when approaching the singularity , e.g. , the singularity inside a black hole and the big bang singularity in frw cosmology .
( c ) nonscalar singularities : the observer experiences unbounded tidal forces just along some world - line curves , e.g. , whimper cosmologies @xcite . based on the energy conditions which imply that the gravity must be attractive , singularities are unavoidable in general relativity . in this
regard , cosmological models that contain nonexotic fluids ( radiation or dust ) , exhibit an initial singularity , the so - called the big bang singularity .
since this fact is unavoidable in general relativity , it is hoped that the quantum theory of gravitation will solve this issue and lead to the avoidance of the singularities in the quantum domain . in spite of the nonexistence of the completely acceptable theory of quantum gravity , it is shown that various approaches that combine the laws of quantum mechanics with the general relativity could be able to partially solve this problem @xcite . in this paper
, we study the hl quantum cosmology in the presence of the chaplygin gas .
note that the chaplygin gas hl classical cosmology has been studied in refs .
bertolami and zarro studied the projectable hl gravity in the context of the minisuperspace model of quantum cosmology for a frw universe without matter @xcite .
also , the hl quantum cosmology in the presence of the perfect fluid has been recently investigated in ref .
@xcite . here
we show that the existence of the chaplygin gas results in a different schrdinger - wheeler - dewitt ( swd ) equation . using the schutz formalism
, we investigate the time evolution of the expectation value of the scale factor in the isotropic and homogenous universe described by the frw spacetime . in the framework of the schutz formalism
, the variable associated to the degrees of freedom of matter plays the role of time and leads to a well - define hilbert space structure .
notice that the proposed approach is a poor approximation of the real dynamics of reducing the brane physics into hoava - lifshitz gravity plus the generalized chaplygin gas .
moreover , the hl gravity model is not fully compatible with general relativity , but it is discussed in order to show that a period of inflation can be obtained . indeed , the hl theory does not exactly recover general relativity at low energy .
however , it mimics general relativity plus dark matter @xcite . on the other hand ,
the nonprojectable hoava - lifshitz gravity is equivalent to einstein - ether theory with a hypersurface orthogonal ether in the ir limit ( see , e.g. , ref .
@xcite for more detail ) . in sec .
[ sec2 ] , we construct chaplygin gas hl quantum cosmology in minisuperspace in terms of velocity potential variables . in sec .
[ sec3 ] , we obtain classical solutions and exhibit the existence of singularities in the classical domain .
then we construct wave packets and find the time evolution of the expectation values of the scale factor to address the existence of singularity - free behavior of the solutions at the quantum level . in sec .
[ sec4 ] , we present our conclusions .
the action for the hl gravity in the presence of chaplygin gas and in the schutz formalism is given by @xmath53 where @xmath54 is defined in eq .
( [ eq9 ] ) and @xmath55 is the action of the chaplygin gas @xcite : @xmath56 since our main purpose is to study frw cosmology , we set @xmath57 due to the symmetry of the frw model .
so , in the rest frame , the four - velocity of the fluid can be written as @xmath58 @xcite which leads to @xmath59 according to ref .
@xcite , thermodynamical relations for the chaplygin gas are given by @xmath60,\nonumber\\ \mu&=&1+\pi+\frac{p}{\rho_0},\nonumber\\ \tau ds & = & d\pi+p\,d\left(\frac{1}{\rho_0}\right),\nonumber\\\label{eq13 } & = & \frac{(1+\pi)^{-\alpha}}{1+\alpha}d\left[(1+\pi)^{1+\alpha}-\frac{a}{\rho_0^{1+\alpha } } \right],\end{aligned}\ ] ] where @xmath61 and @xmath17 are defined as @xmath62 therefore , the equation of state , particle number density , and energy density take the following forms , respectively , @xmath63^{\frac{\alpha}{\alpha+1}},\label{eq16}\\ \rho & = & \left[\frac{1}{a}\left(1-\frac{\,\,\mu^{\frac{\alpha+1}{\alpha}}}{s^{1/\alpha } } \right ) \right]^{\frac{-1}{\,\,1+\alpha}},\label{eq17}\\ \rho_0 & = & \frac{\rho+p}{\mu}.\label{eq18}\end{aligned}\ ] ] the frw metric is @xmath64 where @xmath65 is the scale factor , @xmath66 is the metric for the unit sphere and @xmath67 denotes the open , flat , and closed universes , respectively .
thus , the three - metric is @xmath68 , and @xmath69 .
the ricci curvature tensor @xmath70 and the exterior curvature tensor @xmath43 are given by @xmath71 and the total action takes the form ( in units where @xmath72 ) @xmath73 ^\frac{\alpha}{\alpha+1}\bigg\}.\end{aligned}\ ] ] since the action does not depend on @xmath74 , @xmath75 is indeed the lagrange multiplier .
so , it would not be surprising that the results do not depend on how the spacetime is sliced
. now , the canonical momenta read @xmath76 ^\frac{-1}{\alpha+1},\\ p_{\theta}&=&0,\\ p_s&=&\theta p_\phi , \end{array}\end{aligned}\ ] ] and the super - hamiltonian reads @xcite @xmath77,\nonumber\\ & = & -\frac{p_a^2}{12(3\lambda-1)a}+g_\lambda a^3 -g_c(k)a+\frac{g_r(k)}{a}\nonumber\\\label{eq24 } & & + \frac{g_s(k)}{a^3 } + \left(sp_\phi^{\alpha+1}+aa^{3(\alpha+1)}\right)^\frac{1}{\alpha+1},\end{aligned}\ ] ] where @xmath78 to proceed further , let us consider the following approximation which is valid for the early universe @xcite : @xmath79 so , the super - hamiltonian can be written as @xmath80 now , using the canonical transformations , @xmath81 the super - hamiltonian takes the following form : @xmath82 here , @xmath83 is the only remaining canonical variable associated with the matter .
the classical dynamics of the system is governed by the poisson brackets , namely , @xmath84 . in the quantum domain
, we impose the standard quantization condition on all canonical momenta , i.e. , @xmath85 and @xmath86 .
thus , the swd equation is given by @xmath87 note that the hermicity condition for the hamiltonian implies the following inner product for the wave functions @xcite : @xmath88 for the late times we have @xmath89 so that @xmath90 and the super - hamiltonian takes the form @xmath91 the following canonical transformations , @xmath92 simplify the super - hamiltonian to @xmath93 now , imposing the standard quantization conditions , we obtain the swd equation as @xmath94 demanding that the hamiltonian operator @xmath95 be self - adjoint , the inner product relation takes the form @xcite @xmath96 to investigate the singularity problem , we only present the solutions for the early universe in sec .
[ sec3 ] .
moreover , the wave functions are supposed to obey the boundary conditions @xmath97 where the first condition is called the dewitt boundary condition to avoid the singularity in the quantum domain .
the classical equations of motion are governed by @xmath98 and the constraint equation @xmath99 . by taking @xmath100 ,
the time - independent swd equation takes the form @xmath101 now , we present classical and quantum mechanical solutions for various values of @xmath49s in the following subsections . in the case
@xmath102 where all other @xmath49s are zero , the classical equations of motion in the gauge @xmath103 read @xmath104 after eliminating @xmath105 from eqs .
( [ acl1 ] ) and ( [ acl2 ] ) , we obtain @xmath106 moreover , since @xmath99 , we find @xmath107 for @xmath108 and @xmath109 , the solution of eqs .
( [ acl6 ] ) and ( [ acl7 ] ) reads @xmath110 where @xmath111 and @xmath112 . for this case ,
the time - independent swd equation is @xmath113 so , the solutions are given by @xmath114 where @xmath115 and @xmath116 are bessel functions with order @xmath117 .
thus , the wave function that satisfies the dewitt boundary condition reads @xmath118 notice that , in the context of general relativity and for the frw quantum cosmology with zero spatial curvature , namely , @xmath119 , the order of the bessel functions is @xmath120 @xcite .
now , we construct a wave packet with asymptotic classical behavior upon choosing the appropriate weight function : @xmath121 to this end , we use the new variable @xmath122 and choose the weight function @xmath123 where @xmath124 is an arbitrary positive constant .
so , we have @xmath125 where @xmath126 .
now , using the relation @xmath127 we find the squared integrable wave packet @xmath128 and the expectation value of the scale factor is given by @xmath129 therefore , we obtain @xmath130 since @xmath131 , for @xmath132 ^ 3 $ ] , the universe avoids the singularity in early times and asymptotically tends to the classical solution at late times ( see fig . [ fig : f1 ] ) .
moreover , this model predicts an accelerated universe at the late times .
in fact , eq . ( [ a8 ] ) could be a sign of a nonsingular behavior of the model as well as general relativity quantum cosmology with proper initial conditions .
notice that the nonsingular behavior is due to the dewitt boundary condition , i.e. , @xmath133 , and the gaussian smearing function is chosen to represent the classical - quantum correspondence for large @xmath28 .
the quantum regime of the hl theory of gravity can also provide a suitable framework for the description of the asymptotic darkness of the visible universe @xcite .
note that , various theories of quantum gravity , such as string theory , loop quantum gravity , and black - hole physics all predict the existence of a minimal length scale proportional to the planck length .
the above bouncing behavior could also be related to the presence of a minimal length , namely , @xmath134 ^ 2 \left(\frac{16(3\lambda-1)p_t}{3}\right)^{-\frac{1}{3}}$ ] in agreement with various quantum gravity proposals .
( red line ) and @xmath135 ( blue line ) for @xmath102 in the gauge @xmath103 , i.e. , @xmath131 .
the bouncing behavior of the quantum mechanical solution is apparent at @xmath136.,scaledwidth=45.0% ] in this case , the classical equations of motion in the gauge @xmath138 read @xmath139 after eliminating @xmath105 from eqs .
( [ airycl1 ] ) and ( [ airycl2 ] ) , we obtain @xmath140 also , the constraint @xmath99 leads to @xmath141 for @xmath142 , the solution reads @xmath143 where @xmath144 and @xmath145 . for the quantum solution
, the time - independent swd equation reads @xmath146 using the variable @xmath147 , the above equation can be written as @xmath148 and the solutions are the airy functions : @xmath149 airy functions @xmath150 [ @xmath151 have oscillatory behaviors for @xmath152 ( @xmath153 ) and decrease ( increase ) exponentially for @xmath154 ( @xmath155 ) which shows that @xmath156 is physically unacceptable . therefore , we find @xmath157.\end{aligned}\ ] ] now , the dewitt boundary condition implies @xmath158 where @xmath159 are zeros of the airy function .
therefore , the energy spectrum reads @xmath160 and the time - dependent solutions take the form @xmath161.\end{aligned}\ ] ] notice that , for @xmath162 , the airy s function @xmath150 exhibits an oscillatory behavior for @xmath163 ( @xmath153 ) , whereas for @xmath164 ( @xmath155 ) , it decreases monotonically and for large @xmath165 becomes an exponentially damped function .
therefore , contrary to what is usually expected , eq .
( [ airy7 ] ) represents a classical behavior for small @xmath166 and a quantum behavior for large values of the scale factor .
this is in contrast to the usual expected results .
in fact , detecting quantum gravitational effects in large universes is noticeable which is also observed in the frw , stephani , and kaluza - klein models in the context of general relativity @xcite .
note that for @xmath162 , the solution ( [ airy7 ] ) is squared integrable . in this case
, the dynamics of the classical system in the gauge @xmath103 is governed by @xmath169 so , we obtain @xmath170 also , the constraint @xmath99 leads to @xmath171 for @xmath142 and @xmath167 , the solution reads @xmath172^{\frac{1}{3}},\end{aligned}\ ] ] where @xmath173 . for the quantum solution ,
the time - independent swd equation is @xmath174 for @xmath108 the solution is @xmath175 where @xmath176 and @xmath177 are whittaker functions upon choosing @xmath178 , @xmath179 , and @xmath180 . now the dewitt boundary condition is fulfilled if we set @xmath181 and @xmath133 .
the time - independent swd equation for this case reads @xmath183 the solution for @xmath108 is given by @xmath184 where @xmath185 is the biconfluent heun function @xcite , and @xmath186 , @xmath187 , @xmath188 , @xmath189 , @xmath190 , and @xmath191 .
the above solution with @xmath181 and @xmath192 satisfies the dewitt boundary condition for @xmath193 . in this case , the time - independent swd equation is @xmath194 and the solution for @xmath108 reads @xmath195 where @xmath196 is the triconfluent heun function @xcite , and @xmath197 , @xmath198 , @xmath199 , @xmath200 , and @xmath201 .
note that @xmath196 do not obey the dewitt boundary condition . in the particular case @xmath204 and @xmath205 , the classical dynamics of the system in gauge @xmath206
is governed by @xmath207 so , we have @xmath208 also , using the hamiltonian s constraint @xmath99 , we find @xmath209 thus , the classical solution reads @xmath210 where @xmath211 is the frequency of the oscillation . at the quantum domain ,
the swd equation is @xmath212 where its solution reads @xmath213 here , @xmath214 are hermite functions , @xmath215 , and @xmath216 therefore , the time - dependent squared integrable solution is @xmath217 where @xmath218 is the normalization coefficient and @xmath219 .
the eigenfunctions ( [ eq60 ] ) do not satisfy either of the two boundary conditions ( [ eq32 ] ) and ( [ eq33 ] ) .
indeed , @xmath220 is similar to the stationary quantum wormholes which is defined by hawking .
however , this model does not represent static quantum wormholes which are ruled out by the requirement of the self - adjointness of the hamiltonian @xcite .
now , assuming the minus sign in eq .
( [ eq59en ] ) due to the condition @xmath221 and taking @xmath108 , @xmath182 , and @xmath222 , we construct wave packets upon superimposing the above eigenfunctions ( [ eq60 ] ) : @xmath223 using the initial wave function , @xmath224 the expansion coefficients are given by @xmath225 thus , the time evolution of the expectation value of the scale factor , namely , @xmath226 can be consequently obtained .
note that , the relation between @xmath227 and the cosmic time @xmath28 is given by @xmath228 figure [ fig : f2 ] shows the behavior of @xmath229 for @xmath230 , @xmath231 , @xmath232 , and @xmath233 .
as the figure shows , @xmath229 exhibits an oscillatory and nonvanishing behavior .
this would be a sign that the model is singularity free in the quantum domain .
note that although the eigenfunctions ( [ eq60 ] ) do not satisfy either of the two boundary conditions ( [ eq32 ] ) , ( [ eq33 ] ) , using the initial wave function ( [ eq63 ] ) , the expectation value of the scale factor never tends to the singular point .
this shows that the evolution of the universe is uniquely determined by the initial wave packet , and no boundary condition at @xmath234 is necessary .
a similar behavior is also observed for the perfect fluid hl quantum cosmology @xcite .
the oscillatory behavior of the solution also appears in other quantum cosmological models @xcite . ) with respect to the cosmic time @xmath28 .
we set @xmath230 , @xmath231 , @xmath232 , and @xmath233.,scaledwidth=45.0% ]
in this work , we have studied the hoava - lifshitz quantum cosmology in the presence of the chaplygin gas and in the framework of the schutz formalism . due to the asymmetry of space and time in the hl gravity
, the adm decomposition would be a natural scheme to obtain the hamiltonian of the model .
on the other hand , the chaplygin gas model could describe a transition from a universe filled with dustlike matter to an accelerating expanding stage . as a poor approximation the present approach reduces the brane physics into hoava - lifshitz gravity plus the generalized chaplygin gas .
although the hl gravity is not fully compatible with general relativity , we showed that a period of inflation can be achieved .
in particular , the nonprojectable hoava - lifshitz gravity is equivalent to the einstein - ether theory with a hypersurface orthogonal ether at low energy @xcite . using schutz variational formalism
, we obtained the swd equation where the matter degrees of freedom played the role of time .
we exactly solved the swd equation for various cases . for two cases ,
namely , @xmath102 ( sec . [ seca ] ) and @xmath202 and @xmath203 ( sec . [ secf ] ) , we constructed wave packets upon choosing appropriate initial conditions . in both cases , unlike the classical solutions , the expectation value of the scale factor avoids the singularity in the quantum domain . for @xmath102 the wave function of the universe is given by the bessel functions with order @xmath117 , whereas the order of the bessel functions is @xmath120 for the chaplygin gas frw quantum cosmology with zero spatial curvature ( @xmath119 ) in the context of general relativity . in sec .
[ seca ] , we constructed a wave packet that agrees with the classical solution at late times .
it is worth mentioning that , for this case , the notion of time that arises from schutz formalism coincides with the cosmological time at late times . in sec .
[ secb ] , we found classical behavior for small values of the scale factor and a quantum behavior for its large values . this unexpected result is also observed in the frw , stephani , and kaluza - klein models in the context of general relativity .
the bouncing behavior of the expectation value of the scale factor could be related to the existence of a minimal length scale @xmath235 ^ 2 \left(\frac{16(3\lambda-1)p_t}{3}\right)^{-\frac{1}{3}}$ ] which is suggested by various proposals of quantum gravity . in sec .
[ secf ] , we found that the evolution of the wave function of the universe is uniquely determined by the initial wave packet , i.e. , no boundary condition at @xmath234 is necessary in hl quantum cosmology . among various
presented solutions , solutions ( [ airycl6 ] ) , i.e. , @xmath236 and ( [ whittakercl6 ] ) , i.e. , @xmath237^{\frac{1}{3}}$ ] which show the accelerated expansion of the universe , namely , @xmath238 , could match the inflation scenario .
in particular , solution ( [ airycl6 ] ) tends to the power - law inflation @xmath239 and solution ( [ whittakercl6 ] ) tends to the exponential inflation @xmath240 .
observational data of the cosmic microwave background strongly indicate that the primordial cosmological perturbations have an almost scale - invariant spectrum . in general relativity ( @xmath241 )
, the scale invariance also requires inflation .
indeed , for this case , the scale invariance is nothing but the constancy of the hubble expansion rate @xmath242 which leads to @xmath243 . for general @xmath30 , the amplitude of quantum fluctuations of the scalar field in
hl gravity reads @xcite @xmath244 where @xmath245 is some energy scale .
for @xmath37 it is reduced to @xmath246 , which means that the amplitude of quantum fluctuations does not depend on the hubble expansion rate . in other words ,
the spectrum of cosmological perturbations in hoava - lifshitz gravity with @xmath247 is automatically scale invariant even without inflation .
in fact , the power law expansion @xmath248 with @xmath249 satisfies the condition . in our study , solutions ( [ acl9 ] ) , ( [ airycl6 ] ) , and ( [ whittakercl6 ] ) satisfy this condition .
notice that although solution ( [ acl9 ] ) , i.e. , @xmath250 , is not an accelerating solution , it results in the scale - invariant spectrum in hl gravity .
we would like to thank babak vakili and homa shababi for insightful comments and suggestions .
the authors are also grateful to the referees for giving such constructive comments which considerably improved the quality of the paper .
this research is supported by the iran national science foundation ( insf ) , grant no .
93047987 .
99 supernova search team collaboration , a. g. riess _
et al . _ ,
j. * 116 * , 1009 ( 1998 ) .
supernova cosmology project collaboration , s. perlmutter _
et al . _ ,
j. * 517 * , 565 ( 1999 ) .
j. l. tonry _
_ , astrophys .
j. * 594 * , 1 ( 2003 ) .
d. n. spergel _
et al . _ ,
. j. suppl . *
148 * , 175 ( 2003 ) . c. l. bennett _ et al .
_ , astrophys
. j. suppl . *
148 * , 1 ( 2003 ) .
sdss collaboration , m. tegmark _
et al . _ ,
d * 69 * , 103501 ( 2004 ) .
s. weinberg , rev .
phys . * 61 * , 1 ( 1989 ) .
j. e. peebles and b. ratra , rev .
mod . phys . * 75 * , 559 ( 2003 ) .
m. c. bento , o. bertolami , and a. a. sen , phys .
d * 66 * , 043507 ( 2002 ) . h.
b. benaoum , adv . high energy phys .
* 2012 * , 357802 ( 2012 ) .
h. saadat and b. pourhassan , int .
phys . * 53 * , 1168 ( 2014 ) .
zhai , y - d .
xu , and x - z .
d * 15 * , 1151 ( 2006 ) .
a. y. kamenshchik , u. moschella , and v. pasquier , phys .
b * 511 * , 265 ( 2001 ) .
n. bilic , g. b. tupper , and r. d. viollier , phys .
b * 535 * , 17 ( 2002 ) .
u. debnath , a. banerjee , and s. chakraborty , classical quantum gravity * 21 * , 5609 ( 2004 ) .
h. saadat and b. pourhassan , astrophys .
space sci . * 344 * , 237 ( 2013 ) .
j. naji , b. pourhassan , and a.r .
amani , int .
d * 23 * , 1450020 ( 2013 ) .
b. pourhassan , int . j. mod
d * 22 * , 1350061 ( 2013 ) .
j. sadeghi , b. pourhassan , m. khurshudyan , and h. farahani , int . j. theor .
phys . * 53 * , 911 ( 2014 ) .
m. bouhmadi - lopez and p. v. moniz , phys .
d * 71 * , 063521 ( 2005 ) ; aip conf .
* 736 * , 188 ( 2004 ) .
o. bertolami and v. duvvuri , phys .
b * 640 * , 121 ( 2006 ) .
m. bouhmadi - lopez , c. kiefer , b. sandhofer , and p.v .
moniz , phys .
d * 79 * , 124035 ( 2009 ) .
m. bouhmadi - lopez , p. frazao , and a. b. henriques , phys .
d * 81 * , 063504 ( 2010 ) ; arxiv:1002.4785 .
m. bouhmadi - lopez , p. chen , and y .- w .
liu , phys .
d * 84 * , 023505 ( 2011 ) ; aip conf .
proc . * 1458 * , 327 ( 2012 ) .
m. bouhmadi - lopez , j. morais , and a. b. henriques , phys .
d * 87 * , 103528 ( 2013 ) .
r. jackiw , arxiv : physics/0010042 ; n. ogawa , phys .
d * 62 * , 085023 ( 2000 ) ; a. kamenshchik , u. moschella , and v. pasquier , phys .
b * 487 * , 7 ( 2000 ) ; j.c .
fabris , s.v.b .
goncalves , and p.e .
de souza , gen .
* 34 * , 53 ( 2002 ) . s. del campo and r. herrera , phys .
b * 660 * , 282 ( 2008 ) .
p. vargas moniz , classical quantum gravity * 19 * , l127 ( 2002 ) ; phys .
d * 66 * , 064012 ( 2002 ) ; phys .
d * 66 * , 103501 ( 2002 ) ; h. q. lu , t. harko , and k. s. cheng , int . j. mod
d * 8 * , 625 ( 1999 ) .
m. bouhmadi - lopez and j. a. jimenez madrid , j. cosmol .
astropart . phys .
05 ( 2005 ) 005 ; j
hou and x .- z .
li , phys . lett .
b * 606 * , 7 ( 2005 ) ; s. srivastava , phys .
b * 619 * , 1 ( 2005 ) .
n. arkani - hamed , s. dimopoulos , and g. r. dvali , phys .
b * 429 * , 263 ( 1998 ) .
j. khoury , b.a .
ovrut , p. j. steinhardt , and n. turok , phys .
d * 64 * , 123522 ( 2001 ) .
a. davidson , d. karasik , and y. lederer , classical quantum gravity * 16 * , 1349 ( 1999 ) ; a. davidson and d. karasik , mod .
a * 13 * , 2187 ( 1998 ) ; a. davidson , classical quantum gravity * 16 * , 653 ( 1999 ) .
j. garriga and m. sasaki , phys .
d * 62 * , 043523 ( 2000 ) ; m. bouhmadi and p. f. gonzalez - dyaz , phys .
d * 65 * , 063510 ( 2002 ) ; m. bouhmadi - lopez , p. f. gonzalez - dyaz , and a. zhuk , classical quantum gravity * 19 * , 4863 ( 2002 ) .
k. koyama and j. soda , phys .
b * 483 * , 432 ( 2000 ) .
l. anchordoqui , c. nunez , and k. olsen , j. high energy phys .
10 ( 2000 ) 050 .
s. nojiri and s. d. odintsov , j. cosmol .
astropart . phys .
06 ( 2003 ) 004
. s. s. seahra , h. r. sepangi , and j. ponce de leon , phys .
d * 68 * , 066009 ( 2003 ) .
a. sanyal , arxiv:0305042 .
s. del campo , j. cosmol .
phys . 11 ( 2013 ) 004
. t. barreiro , o. bertolami , and p. torres , phys .
d * 78 * , 043530 ( 2008 ) . since
our study is in the preinflationary phase , the current bounds on the gcg parameters are not relevant for the preinflationary phase .
m. bouhmadi - lopez , pisin chen , y .- c huang , and y .- h lin , phys .
d * 87 * , 103513 ( 2013 ) . c. kiefer and h. d. zeh , phys .
d * 51 * , 4145 ( 1995 ) .
d. giulini , e. joos , c. kiefer , j. kupasch , i .- o .
stamatescu , and h. d. zeh , _ decoherence and the appearance of a classical world in quantum theory _ ( spinger - verlag , berlin , 1996 ) . c. kiefer , _ quantum gravity _
( oxford university press , new york , 2004 ) . l. diosi and c. kiefer , phys . rev .
* 85 * , 3552 ( 2000 ) .
b. f. schutz , phys .
d * 2 * , 2762 ( 1970 ) .
b. f. schutz , phys .
d * 4 * , 3559 ( 1971 ) .
e. gourgoulhon , arxiv : gr - qc/0703035 .
p. horava , j. high energy phys .
03 ( 2009 ) 020 .
p. horava , phys .
d * 79 * , 084008 ( 2009 ) .
p. horava , phys .
102 * , 161301 ( 2009 ) .
p. horava , phys .
b * 694 * , 172 ( 2010 ) .
o. obregon and j. a. preciado , phys .
d * 86 * , 063502 ( 2012 ) .
d. blas , o. pujolas , and s. sibiryakov , j. high energy phys . 04
( 2011 ) 018 .
t. p. sotiriou , m. visser , and s. weinfurtner , j. high energy phys .
10 ( 2009 ) 033 .
j. bellorin and a. restuccia , phys .
d * 84 * , 104037 ( 2011 ) .
a. wang and r maartens , phys .
d * 81 * , 024009 ( 2010 ) .
s. carloni , e. elizalde , and p. j. silva , classical quantum gravity * 27 * , 045004 ( 2010 ) .
s. carloni , e. elizalde , and p. j. silva , classical quantum gravity * 28 * , 195002 ( 2011 ) .
t. zhu , f. -w .
shu , q. wu , and a. wang , phys .
d * 85 * , 044053 ( 2012 ) .
t. p. sotiriou , m. visser , and s. weinfurtner , phys .
lett * 102 * , 251601 ( 2009 ) .
b. vakili and v. kord , gen .
* 45 * 1313 ( 2013 ) .
j. p. m. pitelli and a. saa , phys .
d * 86 * , 063506 ( 2012 ) .
t. m. helliwell and d. a. konkowski , v. arndt , gen .
. gravit . * 35 * , 79 ( 2003 ) .
a. ali , s. dutta , e. n. saridakis , and a. a. sen , gen .
gravit . * 44 * , 657 ( 2012 ) .
b. pourhassan , arxiv:1412.2605 .
b. c. paul , p. thakur , pramana * 81 * , 691 ( 2013 ) .
o. bertolami and c.a.d .
zarro , phys .
d * 84 * , 044042 ( 2011 ) . s. mukohyama , classical quantum gravity * 27 * , 223101 ( 2010 ) .
sotiriou , j. phys .
conf . ser .
* 283 * , 012034 ( 2011 ) . v.
g. lapchinskii and v. a. rubakov , theor .
phys . * 33 * , 1076 ( 1977 ) .
p. pedram , s. jalalzadeh , and s. s. gousheh , int .
. phys . * 46 * , 3201 ( 2007 ) .
p. pedram and s. jalalzadeh , phys . lett .
b * 659 * , 6 ( 2008 ) .
e. elizalde and p. j. silva , j. high energy phys . 06 ( 2012 ) 049 .
p. pedram , s. jalalzadeh , and s.s .
gousheh , classical quantum gravity * 24 * , ( 2007 ) .
n.a . lemos and f.g .
alvarenga , gen .
. gravit . * 31 * , 1743 ( 1999 ) .
r. colistete jr . , j.c .
fabris , and n. pinto - neto , phys .
d * 57 * , 4707 ( 1998 ) .
a. m. ishkhanyan , arxiv:1412.0654
. f. g. alvarenga , j. c. fabris , n. a. lemos , and g. a. monerat , gen .
. gravit . * 34 * , 651 ( 2002 ) .
monerat , e.v.c .
silva , g. oliveira - neto , l.g.f .
filho , and n.a .
lemos , phys .
d * 73 * , 044022 ( 2006 ) . | in this paper , we study the chaplygin gas hoava - lifshitz quantum cosmology . using schutz formalism and arnowitt - deser - misner decomposition ,
we obtain the corresponding schrdinger - wheeler - dewitt equation .
we obtain exact classical and quantum mechanical solutions and construct wave packets to study the time evolution of the expectation value of the scale factor for two cases .
we show that unlike classical solutions and upon choosing appropriate initial conditions , the expectation value of the scale factor never tends to the singular point which exhibits the singularity - free behavior of the solutions in the quantum domain . |
let @xmath0 be a standard one - dimensional brownian motion , @xmath1 be an integer , and let @xmath2 be a symmetric element of @xmath3
. denote by @xmath4 the @xmath5-tuple wiener - it integral of @xmath2 with respect to @xmath6 .
it is well known that multiple wiener - it integrals of different orders are uncorrelated but not necessarily independent . in an important paper @xcite , stnel and
zakai gave the following characterization of the independence of multiple wiener - it integrals .
[ uzthmintro ] let @xmath7 be integers and let @xmath8 and @xmath9 be symmetric . then , random variables @xmath10 and @xmath11 are independent if and only if @xmath12 rosiski and samorodnitsky @xcite observed that multiple wiener - it integrals are independent if and only if their squares are uncorrelated : @xmath13 this condition can be viewed as a generalization of the usual covariance criterion for the independence of jointly gaussian random variables ( the case of @xmath14 ) . in the seminal paper @xcite , nualart and peccati discovered the following surprising central limit theorem .
[ npthmintro ] let @xmath15 , where @xmath16 is fixed and @xmath17 are symmetric .
assume also that @xmath18= 1 $ ] for all @xmath19 .
then convergence in distribution of @xmath20 to the standard normal law is equivalent to convergence of the fourth moment .
that is , as @xmath21 , @xmath22\to 3.\ ] ] shortly afterwards , peccati and tudor @xcite established a multidimensional extension of theorem [ npthmintro ] .
since the publication of these two important papers , many improvements and developments on this theme have been considered . in particular , nourdin and peccati
@xcite extended theorem [ npthmintro ] to the case when the limit of @xmath23 s is a centered gamma distributed random variable .
we refer the reader to the book @xcite for further information and details of the above results .
heuristic argument linking theorem [ uzthmintro ] and theorem [ npthmintro ] was given by rosiski @xcite , while addressing a question of albert shiryaev .
namely , let @xmath24 and @xmath25 be two i.i.d
. centered random variables with fourth moment and unit variance .
the link comes via a simple formula @xmath26 - 3,\ ] ] criterion , as well as the celebrated bernstein s theorem that asserts that @xmath24 and @xmath25 are gaussian if and only if @xmath27 and @xmath28 are independent .
a rigorous argument to carry through this idea is based on a characterization of the asymptotic independence of multiple wiener - it integrals , which is much more difficult to handle than the plain independence , and may also be of an independent interest .
the covariance between the squares of multiple wiener - it integrals plays the pivotal role in this characterization . at this point
we should also mention an extension of to the multivariate setting .
let @xmath29 be a finite set and @xmath30 be a sequence of non - negative integers .
let @xmath31 be a multiple wiener - it integral of order @xmath32 , @xmath33 .
consider a partition of @xmath29 into disjoint blocks @xmath34 , so that @xmath35 , and the resulting random vectors @xmath36 , @xmath37
. then @xmath38 the proof of this criterion is similar to the proof of in @xcite . in this paper in theorem [ mainblock ]
we establish an asymptotic version of characterizing the asymptotic moment - independence between blocks of multiple wiener - it integrals . as a consequence of this result ,
we deduce the fourth moment theorem of nualart and peccati @xcite in theorem [ nupec ] , its multidimensional extension due to peccati and tudor @xcite in theorem [ pectud ] , and some neat estimates on the speed of convergence in theorem [ pectud - bd ] . furthermore
, we obtain new multidimensional extension of a theorem of nourdin and peccati @xcite in theorem [ t : noupec ] , and give another new result on the bivariate convergence of vectors consisting of multiple wiener - it integrals in theorem [ t : bivariateblock ] .
proposition [ bm - dmt ] applies theorem [ t : bivariateblock ] to establish the limit process for functions of short and long range dependent stationary gaussian time series in the spirit of the celebrated breuer - major @xcite and dobrushin - major - taqqu @xcite theorems . in theorem [ thmmoo ]
we establish the asymptotic moment - independence for discrete non - gaussian chaoses using some techniques of mossel , odonnel and oleszkiewicz @xcite .
the paper is organized as follows . in section [ s : pre ]
we list some basic facts from gaussian analysis and prove some lemmas needed in the present work .
in particular , we establish lemma [ l : cs ] , which a version of the cauchy - schwarz inequality well suited to deal with contractions of functions , see .
it is used in the proof of the main result , theorem [ mainblock ] .
section [ s : main ] is devoted to the main results on the asymptotic independence . section [ s : apps ] gives some immediate consequences and related applications of the main result .
section [ s : f - apps ] provides further applications to the study of short and long range dependent stochastic processes and multilinear random forms in non - gaussian random variables .
we will give here some basic elements of gaussian analysis that are in the foundations of the present work .
the reader is referred to the books @xcite for further details and ommited proofs .
let @xmath39 be a real separable hilbert space .
for any @xmath1 let @xmath40 be the @xmath5th tensor product of @xmath39 and denote by @xmath41 the associated @xmath5th symmetric tensor product .
we write @xmath42 to indicate an isonormal gaussian process over @xmath39 , defined on some probability space @xmath43 .
this means that @xmath44 is a centered gaussian family , whose covariance is given in terms of the inner product of @xmath39 by @xmath45 = \langle h , g\rangle _ { \eufrak h}$ ] .
we also assume that @xmath46 is generated by @xmath44 . for every @xmath1 , let @xmath47 be the @xmath5th wiener chaos of @xmath44 , that is , the closed linear subspace of @xmath48 generated by the random variables of the type @xmath49 , where @xmath50 is the @xmath5th hermite polynomial defined as @xmath51 we write by convention @xmath52 . for any @xmath1
, the mapping @xmath53 can be extended to a linear isometry between the symmetric tensor product @xmath41 equipped with the modified norm @xmath54 and the @xmath5th wiener chaos @xmath47 .
for @xmath55 we write @xmath56 , @xmath57 .
it is well known ( wiener chaos expansion ) that @xmath48 can be decomposed into the infinite orthogonal sum of the spaces @xmath47 . therefore
, any square integrable random variable @xmath58 admits the following chaotic expansion @xmath59 where @xmath60 $ ] , and the @xmath61 , @xmath1 , are uniquely determined by @xmath24 . for every @xmath62
we denote by @xmath63 the orthogonal projection operator on the @xmath5th wiener chaos . in particular , if @xmath58 is as in ( [ e ] ) , then @xmath64 for every @xmath62 .
let @xmath65 be a complete orthonormal system in @xmath39 .
given @xmath66 and @xmath67 , for every @xmath68 , the _ contraction _ of @xmath2 and @xmath69 of order @xmath70 is the element of @xmath71 defined by @xmath72 notice that @xmath73 is not necessarily symmetric : we denote its symmetrization by @xmath74 .
moreover , @xmath75 equals the tensor product of @xmath2 and @xmath69 while , for @xmath76 , @xmath77 . in the particular case where @xmath78 , where @xmath79 is a measurable space and @xmath80 is a @xmath81-finite and non - atomic measure
, one has that @xmath82 is the space of symmetric and square integrable functions on @xmath83 .
moreover , for every @xmath84 , @xmath85 coincides with the @xmath5-tuple wiener - it integral of @xmath2 . in this case , ( [ v2 ] ) can be written as @xmath86 we have @xmath87 where @xmath88 ( @xmath89 , respectively ) stands for inner product ( the norm , respectively ) in an appropriate tensor product space @xmath90 .
also , the following _ multiplication formula _ holds : if @xmath91 and @xmath92 , then @xmath93 where @xmath94 denotes the symmetrization of @xmath95 .
we conclude these preliminaries by three useful lemmas , that will be needed throughout the sequel .
[ usefullemma1 ]
-0.2 in * multiple wiener - it integral has all moments satisfying the following hypercontractivity - type inequality @xmath96^{1/r } \le ( r-1)^{p/2 } \big [ e |i_p(f)|^2\big]^{1/2 } , \quad r{\geqslant}2.\ ] ] * if a sequence of distributions of @xmath97 is tight , then @xmath98 _ proof_. ( i ) inequality ( [ hc ] ) is well known and corresponds e.g. to ( * ? ? ?
* corollary 2.8.14 ) .
\(ii ) combining ( [ hc ] ) for @xmath99 with paley s inequality we get for every @xmath100 @xmath101 by the assumption , there is an @xmath102 such that @xmath103 , @xmath104 . by with @xmath105 and all @xmath19
, we have @xmath106 as a consequence , @xmath107 . applying we conclude .
[ technical ] -02 in 1 .
let @xmath7 , @xmath91 and @xmath92 .
then @xmath108 2 .
let @xmath1 and @xmath109 .
then @xmath110 3 .
let @xmath1 , @xmath111 and @xmath112 .
we have @xmath113 _ proof_. without loss of generality , we suppose throughout the proof that @xmath114 is equal to @xmath115 , where @xmath79 is a measurable space and @xmath116 is a @xmath117-finite measure without atoms .
\(1 ) let @xmath117 be a permutation of @xmath118 ( this fact is written in symbols as @xmath119 ) .
if @xmath120 denotes the cardinality of @xmath121 , then it is readily checked that @xmath70 is also the cardinality of @xmath122 and that @xmath123 moreover , for any fixed @xmath120 , there are @xmath124 permutations @xmath119 such that @xmath125 .
( indeed , such a permutation is completely determined by the choice of : @xmath126 @xmath70 distinct elements @xmath127 of @xmath128 ; @xmath129 @xmath130 distinct elements @xmath131 of @xmath132 ; @xmath133 a bijection between @xmath132 and @xmath134 ; @xmath135 a bijection between @xmath128 and @xmath136 .
) now , observe that the symmetrization of @xmath137 is given by @xmath138 therefore , using ( [ ctr ] ) , we can write @xmath139 and ( [ sym - norm ] ) follows
. \(2 ) we proceed analogously . indeed
, we have @xmath140 from which we deduce ( [ azerty ] ) .
\(3 ) we have @xmath141 and @xmath142 so that @xmath143 since @xmath144 , the desired conclusion ( [ azertiop ] ) follows .
[ l : cs ] assume that @xmath145 , where @xmath79 is a measurable space equipped with a @xmath117-finite measure @xmath116 . for any integer @xmath146 , put @xmath147 = \{1, ... ,m\}$ ] .
also , for every element @xmath148 and every nonempty set @xmath149 $ ] , let @xmath150 denote the element of @xmath151 ( where @xmath152 is the cardinality of @xmath153 ) obtained by deleting from @xmath154 the entries with index not contained in @xmath153 .
( for instance , if @xmath155 and @xmath156 , then @xmath157 . )
let 1 .
@xmath158 be integers , and let @xmath159 be nonempty subsets of @xmath160 $ ] such that each element of @xmath160 $ ] appears in exactly two of the @xmath161 s ( this implies that @xmath162 $ ] and @xmath163 ) ; 2 .
let @xmath164 be functions such that @xmath165 for every @xmath166 ( in particular , each @xmath167 is a function of @xmath168 variables ) .
then @xmath169})\big| \leq \prod_{i=1}^q \|h_i\|_{l^2(\mu^{|c_i|})}.\ ] ] moreover , if @xmath170 for some @xmath171 , then @xmath172 } ) \big|\leq \|h_j \otimes_{c_0 } h_k\|_{l^2(\mu^{|c_j \triangle c_k| } ) } \prod_{i\ne j , k}^q \|h_i\|_{l^2(\mu^{|c_i|})},\ ] ] where @xmath173 ( notice that @xmath174 when @xmath175 and @xmath176 are symmetric . )
_ proof_. in the case @xmath177 , is just the cauchy - schwarz inequality and is an equality .
assume that
hold for at most @xmath178 functions and proceed by induction . among the sets @xmath179 at least two , say @xmath180 and @xmath181 , have nonempty intersection .
set @xmath182 , as above .
since @xmath183 does not have common elements with @xmath161 for all @xmath184 , by fubini s theorem @xmath185 } ) = \int_{a^{c-|c_0| } } h_j
\otimes_{c_0 } h_k({\bf z}_{c_j \triangle c_k } ) \prod_{i\ne j , k}^q h_i({\bf z}_{c_i } ) \ , \mu^{c-|c_0|}(d{\bf z}_{[c ] \setminus c_0}).\ ] ] observe that every element of @xmath160 \setminus c_0 $ ] belongs to exactly two of the @xmath178 sets : @xmath186 , @xmath161 , @xmath187 .
therefore , by the induction assumption , implies , provided @xmath188 . when @xmath189 , we have @xmath190 and follows from applied to the product of @xmath191 functions in .
this proves , which in turn yields by the cauchy - schwarz inequality .
the proof is complete .
the following theorem characterizes moment - independence of limits of multiple wiener - it integrals .
[ main ] let @xmath192 , and let @xmath193 be positive integers . consider vectors @xmath194 with @xmath195 .
assume that for some random vector @xmath196 , @xmath197 then @xmath198 s admit moments of all orders and the following three conditions are equivalent : 1 .
@xmath199 are moment - independent , that is , @xmath200 = e[u_1^{k_1}]\ldots e[u_d^{k_d}]$ ] for all @xmath201 ; 2 .
@xmath202 for all @xmath203 ; 3 .
@xmath204 for all @xmath203 and all @xmath205 ; moreover , if the distribution of each @xmath198 is determined by its moments , then ( a ) is equivalent to that 1 .
@xmath199 are independent .
[ rm1 ] 1 .
theorem [ main ] raises a question whether the moment - independence implies the usual independence under weaker conditions than the determinacy of the marginals .
( recall that a random variable having all moments is said to be determinate if any other random variable with the same moments has the same distribution . )
the answer is negative in general , see ( * ? ? ? * theorem 5 ) .
2 . assume that @xmath206 ( for simplicity ) .
in this case , ( @xmath207 ) becomes @xmath208 for all @xmath209 . in view of theorem [ uzthmintro ] of stnel and zakai
, one may expect that ( @xmath207 ) could be replaced by a weaker condition ( @xmath210 ) : @xmath211 .
+ however , the latter is false . to see it ,
consider a sequence @xmath212 such that @xmath213 and @xmath214 . by theorem [ nupec ] below , @xmath215 .
putting @xmath216 , we observe that ( @xmath210 ) holds but ( @xmath217 ) does not , as @xmath218 .
3 . taking into account that assumptions @xmath219 and @xmath220 of theorem [ nupec ] are equivalent , it is natural to wonder whether assumption ( @xmath207 ) of theorem [ main ] is equivalent to its symmetrized version : @xmath221 the answer is negative in general , as is shown by the following counterexample .
let @xmath222 ^ 2\to { \mathbb{r}}$ ] be symmetric functions given by @xmath223\\ 1&\quad\mbox{elsewhere } \end{array}\right .
\quad\mbox{and}\quad f_2(s , t)=\left\{\begin{array}{cl } -1&\quad s , t\in(1/2,1]\\ 1&\quad\mbox{elsewhere}. \end{array}\right.\ ] ] then @xmath224 and @xmath225 $ and $ t\in(1/2,1]$}\\ 1&\quad\mbox{if $ t\in[0,1/2]$ and $ s\in(1/2,1]$}\\ 0&\quad\mbox{elsewhere } , \end{array } \right.\ ] ] so that @xmath226 and @xmath227 .
the condition of moment - independence , ( @xmath217 ) of theorem [ main ] , can also be stated in terms of cumulants .
recall that the joint cumulant of random variables @xmath228 is defined by @xmath229}_{\big| t_1=0,\ldots , t_m=0},\ ] ] provided @xmath230 . when all @xmath231 are equal to @xmath44 , then @xmath232 , the usual @xmath233th cumulant of @xmath44 , see @xcite .
then theorem [ main](@xmath217 ) is equivalent to * _ for all integers @xmath234 , @xmath235 , and @xmath236 _
@xmath237 theorem [ main ] was proved in the first version of this paper @xcite .
our proof of the crucial implication @xmath238 involved tedious combinatorial considerations .
we are thankful to an anonymous referee who suggested a shorter and more transparent line of proof using malliavin calculus .
it significantly reduced the amount of combinatorial arguments of the original version but requires some basic facts from malliavin calculus .
we incorporated referee s suggestions and approach into the proof of a more general theorem [ mainblock ] . even though theorem [ main ] becomes a special case of theorem [ mainblock ] ( see corollary [ marginalblock ] ) , we keep its original statement for a convenient reference . for each @xmath239 ,
let @xmath240 be a family of real - valued random variables indexed by a finite set @xmath29 .
consider a partition of @xmath29 into disjoint blocks @xmath34 , so that @xmath35 .
we say that vectors @xmath241 , @xmath37 are asymptotically moment - independent if each @xmath242 admits moments of all orders and for any sequence @xmath243 of non - negative integers , @xmath244 -\prod_{k=1}^d e\big[\prod_{i\in i_k } f_{i , n}^{\ell_i}\big ] \big\}=0.\ ] ] the next theorem characterizes the asymptotic moment - independence between blocks of multiple wiener - it integrals .
[ mainblock ] let @xmath29 be a finite set and @xmath30 be a sequence of non - negative integers .
for each @xmath239 , let @xmath240 be a family of multiple wiener - it integrals , where @xmath245 with @xmath195 .
assume that for every @xmath33 @xmath246 < \infty.\ ] ] given a partition of @xmath29 into disjoint blocks @xmath34 , the following conditions are equivalent : 1 .
random vectors @xmath241
, @xmath37 are asymptotically moment - independent ; 2 .
@xmath247 for every @xmath248 from different blocks ; 3 .
@xmath249 for every @xmath248 from different blocks and @xmath205 .
_ proof _ : the implication @xmath250 is obvious . to show @xmath251 , fix @xmath248 belonging to different blocks . by we have @xmath252 which yields @xmath253 = \sum_{r=0}^{q_i\wedge q_j } r!^2\binom{q_i}{r}^2\binom{q_j}{r}^2(q_i+q_j-2r ) !
\|f_{i , n}\widetilde{\otimes}_r f_{j , n}\|^2.\ ] ] moreover , @xmath254e[f_{j , n}^2 ] = q_i!q_j!\|f_{i , n}\|^2\|f_{j , n}\|^2.\ ] ] applying ( [ sym - norm ] ) to the second equality below , we evaluate @xmath255 as follows : @xmath256 this bound yields the desired conclusion .
now we will prove @xmath257 .
we need to show for fixed @xmath258 . writing @xmath259 as @xmath260 and enlarging @xmath29 and @xmath34 s accordingly , we may and do assume that all @xmath261
. we will prove ( [ blocks ] ) by induction on @xmath262 .
the formula holds when @xmath263 or 1 .
therefore , take @xmath264 and suppose that ( [ blocks ] ) holds whenever @xmath265 .
fix @xmath266 and set @xmath267 assume that @xmath268 , otherwise the inductive step follows immediately .
let @xmath269 denote the divergence operator in the sense of malliavin calculus and let @xmath270 be the malliavin derivative , see ( * ? ? ?
- 1.3 ) . using the duality relation (
1.3.1(ii ) ) and the product rule for the malliavin derivative ( * ? ? ?
* theorem 3.4 ) we get @xmath271 & = e\big[i_{q_{i_1}}(f_{i_1,n})x_ny_n\big ] = e\big[\delta(i_{q_{i_1}-1}(f_{i_1,n}))x_ny_n\big]\\ & = e\big[i_{q_{i_1}-1}(f_{i_1,n})\otimes_1 d(x_ny_n)\big]\\ & = e\big [ y_n \,i_{q_{i_1}-1}(f_{i_1,n})\otimes_1 dx_n \big ] + e\big [ x_n \,i_{q_{i_1}-1}(f_{i_1,n})\otimes_1 dy_n \big ] \\ & = a_n+b_n.\end{aligned}\ ] ] first we consider @xmath272 . using the product rule for @xmath273 we obtain @xmath274\\ & = \sum_{j\in i \setminus i_1 } q_j e \big
[ i_{q_{i_1}-1}(f_{i_1,n})\otimes_1 i_{q_j-1}(f_{j , n } ) \prod_{i\in i \setminus \{i_1 , j\ } } f_{i , n } \big].\end{aligned}\ ] ] by the multiplication formula ( [ multiplication ] ) we have @xmath275 since @xmath276 and @xmath277 belong to different blocks , condition ( c ) of the theorem applied to the above expansion yields that @xmath278 converges to zero in @xmath279 . combining this with and lemma [ usefullemma1 ]
we infer that @xmath280 .
now we consider @xmath281 . if @xmath282 , then @xmath283 by convention and so @xmath284 .
hence @xmath285 -e\big[f_{i_1,n}\big ] \prod_{k=2}^d e\big[\prod_{i\in i_k } f_{i , n}\big ] \big\}=\lim_{n \to \infty } b_n=0.\ ] ] therefore , we now assume that @xmath286 .
write @xmath287 $ ] , where @xmath288 thus @xmath281 is a linear combination of the terms @xmath289,\ ] ] where @xmath290 , @xmath291 , @xmath292 .
the term under expectation is a product of multiple integrals of orders summing to @xmath293 .
therefore , the induction hypothesis applies provided @xmath294 for all @xmath295 with @xmath296 and all @xmath297 .
suppose that holds .
then by the induction hypothesis @xmath298 e[y_n]\big\ } & = 0 .
\end{aligned}\ ] ] moreover , @xmath299=e\big[i_{q_{i_1}-1}(f_{i_1,n})\otimes_1 dx_n\big]=e[i_{q_{i_1}}(f_{i_1,n})x_n]= e\big[\prod_{i\in i_1 } f_{i , n}\big].\ ] ] hence , by the induction hypothesis applied to @xmath300 and the uniform boundedness of all moments of @xmath242 , we get @xmath301 -\prod_{k=1}^d e\big[\prod_{i\in i_k } f_{i , n}\big ] \big\ } = \lim_{n\to\infty } \big\{a_n - e[z_n ] e[y_n]\big\ } = 0 .
\end{aligned}\ ] ] it remains to show .
to this aim we will describe the structure of the terms under the limit . without loss of generality
we may assume that @xmath302 , where @xmath79 is a measurable space and @xmath116 is a @xmath117-finite measure without atoms .
recall notation of lemma [ l : cs ] .
for every integer @xmath146 , put @xmath147 = \{1, ... ,m\}$ ] .
also , for every element @xmath148 and every nonempty set @xmath149 $ ] , we denote by @xmath150 the element of @xmath151 ( where @xmath152 is the cardinality of @xmath153 ) obtained by deleting from @xmath154 the entries with index not contained in @xmath153 .
( for instance , if @xmath155 and @xmath156 , then @xmath157 . )
observe that @xmath303 is a linear combination of functions @xmath304 , @xmath305 obtained as follows .
set @xmath306 and @xmath307 , so that @xmath308 .
choose @xmath309 $ ] such that @xmath310 , @xmath311 and @xmath312 , and then choose @xmath313 $ ] such that @xmath314 and @xmath315 .
it follows that @xmath316 $ ] and @xmath317 .
therefore , each element of @xmath147 $ ] belongs exactly to one or two @xmath318 s .
let @xmath319 : j \ \text{belongs to two sets } \
b_i \}\ ] ] and put @xmath320 \setminus j$ ]
. then @xmath303 is a linear combination of functions of the form @xmath321 where the summation goes over all choices @xmath322 under the constraint that the sets @xmath323 and @xmath324 are fixed .
this constraint makes @xmath325 unique , @xmath326 .
let @xmath327 , @xmath328 and notice that either @xmath329 or @xmath330 since @xmath331 .
suppose @xmath332 , the other case is identical .
applying lemma [ l : cs ] with @xmath333 fixed we get @xmath334 and @xmath335 make a disjoint partition of @xmath325 , and additional integration with respect to @xmath333 yields @xmath336 as @xmath337 . this yields and completes the proof of theorem [ mainblock ] .
[ rm2 ] condition ( @xmath338 ) of theorem [ mainblock ] is equivalent to * _ for every @xmath339 _
@xmath340 where @xmath341 denotes the euclidean norms in @xmath342 and @xmath343 respectively . indeed , condition ( * b * ) of theorem [ mainblock ] implies ( * b * ) and the converse follows from @xmath344 as the squares of multiple wiener - it integrals are non - negatively correlated , cf . .
the following corollary is useful to deduce the joint convergence in law from the convergence of marginals .
it is stated for random vectors , as is theorem [ mainblock ] , but it obviously applies in the setting of theorem [ main ] when all vectors are one - dimensional .
[ marginalblock ] under notation of theorem [ mainblock ] , let @xmath345 be a random vector such that * @xmath346 as @xmath347 , for each @xmath348 ; * vectors @xmath349 , @xmath37 are independent ; * condition @xmath350 or @xmath351 of theorem [ mainblock ] holds [ equivalently , @xmath352 or @xmath219 of theorem [ main ] when all @xmath34 are singletons ] ; * @xmath353 is determined by its moments for each @xmath354 .
then the joint convergence holds , @xmath355 _ proof : _ by ( i ) the sequence @xmath356 is tight .
let @xmath357 be a random vector such that @xmath358 as @xmath359 along a subsequence . from lemma [ usefullemma1](ii )
we infer that condition of theorem [ mainblock ] is satisfied .
it follows that each @xmath360 has all moments and @xmath361 for each @xmath348 . by ( iv ) , the laws of vectors @xmath362 and @xmath363 are determined by their joint moments , respectively , see ( * ? ? ?
* theorem 3 ) . under the assumption ( iii ) , the vectors @xmath241 , @xmath37 are asymptotically moment independent .
hence , for any sequence @xmath243 of non - negative integers , @xmath364 - e\big[\prod_{i\in i } u_{i}^{\ell_i}\big ] & = e\big[\prod_{i\in i } v_{i}^{\ell_i}\big ] - \prod_{k=1}^d e\big[\prod_{i\in i_k } u_{i}^{\ell_i}\big ] \\ & = \lim_{n_j\to\infty } \big\ { e\big[\prod_{i\in i } f_{i , n_j}^{\ell_i}\big ] -\prod_{k=1}^d e\big[\prod_{i\in i_k } f_{i , n_j}^{\ell_i}\big ] \big\}=0.\end{aligned}\ ] ] thus @xmath365 .
we can give a short proof of the difficult and surprising part implication @xmath366 of the fourth moment theorem of nualart and peccati @xcite , that we restate here for a convenience .
[ nupec ] let @xmath20 be a sequence of the form @xmath15 , where @xmath16 is fixed and @xmath367 .
assume moreover that @xmath18=q!\|f_n\|^2=1 $ ] for all @xmath19 .
then , as @xmath21 , the following four conditions are equivalent : 1
. @xmath368 ; 2 .
@xmath369\to 3 $ ] ; 3 .
@xmath370 for all @xmath371 ; 4 .
@xmath372 for all @xmath371 .
_ proof of _ @xmath366 .
assume @xmath352 .
since the sequence @xmath20 is bounded in @xmath373 by the assumption , it is relatively compact in law . without loss of generality
we may assume that @xmath374 and need to show that @xmath375 .
let @xmath376 be an independent copy of @xmath23 of the form @xmath377 with @xmath378 .
this can easily be done by extending the underlying isonormal process to the direct sum @xmath379 .
we then have @xmath380 as @xmath21 , where @xmath381 stands for an independent copy of @xmath382 .
since @xmath383 = e[f_{n}^4 ] - 3 \to 0,\ ] ] @xmath384 and @xmath385 are moment - independent .
( if they were independent , the classical bernstein theorem would conclude the proof . )
however , in our case condition ( @xmath386 ) in says that @xmath387 taking @xmath388 we get @xmath389 where we used the multilinearity of @xmath390 and the fact that @xmath382 and @xmath381 are i.i.d . since @xmath391 , @xmath392 , and @xmath393 for @xmath394
, we infer that @xmath375 . applying our approach
, one can add a further equivalent condition to a result of peccati and tudor @xcite .
as such , theorem [ pectud ] turns out to be the exact multivariate equivalent of theorem [ nupec ] .
[ pectud ] let @xmath192 , and let @xmath395 be positive integers .
consider vectors @xmath396 with @xmath195 .
assume that , for @xmath397 , as @xmath21 , @xmath398 let @xmath399 be a centered gaussian random vector with the covariance matrix @xmath400 .
then the following two conditions are equivalent ( @xmath21 ) : 1 .
@xmath401 ; 2 .
@xmath402\to e\big[\|n\|^4\big]$ ] ; where @xmath89 denotes the euclidean norm in @xmath403 . _
proof_. only @xmath404 has to be shown .
assume @xmath405 .
as in the proof of theorem [ nupec ] , we may assume that @xmath374 and must show that @xmath406 .
let @xmath407 be an independent copy of @xmath23 of the form @xmath408 .
observe that @xmath409 - \left(e[{\left\vert f_{n}\right\vert}^2]\right)^2 - 2 \sum_{i , j=1}^d \mathrm{cov}(f_{i , n } , f_{j , n})^2 .
\nonumber \end{aligned}\ ] ] using this identity for @xmath399 and @xmath410 in place of @xmath23 and @xmath376 , where @xmath410 is an independent copy of @xmath399 , we get @xmath411 = \sum_{i , j=1}^d \big ( \sigma_{ii } \sigma_{jj } + 2 \sigma_{ij}^2 \big).\ ] ] hence @xmath412 -e[{\left\vert n\right\vert}^4 ] \\ & + \sum_{i , j=1}^d \big [ \sigma_{ii } \sigma_{jj } + 2 \sigma_{ij}^2 - \mathrm{var}(f_{i , n})\mathrm{var}(f_{j , n } ) - 2 \mathrm{cov}(f_{i , n } , f_{j , n})^2\big ] \to 0.\end{aligned}\ ] ] by remark [ rm2 ] , @xmath413 and @xmath414 are asymptotically moment - independent . since one
- dimensional projections of @xmath413 and @xmath414 are also asymptotically moment - independent , we can proceed by cumulants as above to determine the normality of @xmath382 .
the following result associates neat estimates to theorem [ pectud ] .
[ pectud - bd ] consider a vector @xmath415 with @xmath416 , and let @xmath400 be the covariance matrix of @xmath24 , @xmath417 $ ] .
let @xmath399 be the associated gaussian random vector , @xmath418 . 1 .
assume that @xmath419 is invertible .
then , for any lipschitz function @xmath420 we have @xmath421 - e[h(n ) ] \big| \leq \sqrt{d}\,\|\sigma\|_{op}^{1/2}\|\sigma^{-1}\|_{op}\|h\|_{lip } \sqrt{e\|f\|^4-e\|n\|^4},\ ] ] where @xmath422 denotes the operator norm of a matrix and @xmath423 .
2 . for any @xmath424-function @xmath420
we have @xmath421 - e[h(n ) ] \big| \leq \frac12\|h''\|_\infty \sqrt{e\|f\|^4-e\|n\|^4},\ ] ] where @xmath425 . _ proof _ : the proof is divided into three steps
. _ step 1 _ : recall that for a lipschitz function @xmath420 ( * ? ? ? * theorem 6.1.1 ) yields @xmath421 - e[h(n ) ] \big|
\leq \sqrt{d}\,\|\sigma\|_{op}^{1/2}\|\sigma^{-1}\|_{op}\|h\|_{lip}\sqrt{\sum_{i , j=1}^d e\left\ { \big ( \sigma_{ij}-\frac{1}{q_j}\langle df_i , df_j\rangle \big)^2 \right\}},\ ] ] while for a @xmath424-function with bounded hessian ( * ? ? ?
* theorem 6.1.2 ) gives @xmath421 - e[h(n ) ] \big| \leq \frac12\|h''\|_\infty
\sqrt{\sum_{i , j=1}^de\left\ { \big ( \sigma_{ij}-\frac{1}{q_j}\langle df_i , df_j\rangle \big)^2 \right\}}.\ ] ] _ step 2 _ : we claim that for any @xmath397 , @xmath426 indeed , by ( * ? ? ?
* identity ( 6.2.4 ) ) and the fact that @xmath427 if @xmath428 , we have @xmath429 on the other hand , from ( [ cov ] ) we have @xmath430 the claim follows immediately .
_ step 3 _ : applying we get @xmath431 - \sigma_{ii } \sigma_{jj } - 2 \sigma_{ij}^2 \big)\\ & = \sum_{i , j=1}^d \big\{{\rm cov}(f_i^2,f_j^2 ) -2\sigma_{ij}^2\big\}.\end{aligned}\ ] ] combining steps 1 - 3 gives the desired conclusion .
here we will prove a multivariate extension of a result of nourdin and peccati @xcite .
such an extension was an open problem as far as we know . in what follows
, @xmath433 will denote a random variable with the centered @xmath432 distribution having @xmath434 degrees of freedom .
when @xmath435 is an integer , then @xmath436 , where @xmath437 are i.i.d .
standard normal random variables . in general
, @xmath433 is a centered gamma random variable with a shape parameter @xmath438 and scale parameter @xmath439 .
nourdin and peccati @xcite established the following theorem .
[ noupec ] fix @xmath434 and let @xmath433 be as above .
let @xmath16 be an even integer , and let @xmath15 be such that @xmath440 = e[g(\nu)^2]=2\nu$ ]
. set @xmath441 ^ 3 \left[q!\right]^{-2}$ ] .
then , the following four assertions are equivalent , as @xmath347 : 1 .
@xmath442 ; 2 . @xmath369 -12e[f_n^3 ] \to e[g(\nu)^4 ] -12e[g(\nu)^3 ] = 12\nu^2 - 48\nu$ ] ; 3 . @xmath443 , and @xmath444 for every @xmath445 such that @xmath446 ; 4 .
@xmath443 , and @xmath447 for every @xmath445 such that @xmath446 .
the following is our multivariate extension of this theorem .
[ t : noupec ] let @xmath192 , let @xmath448 be positive reals , and let @xmath449 be _ even _ integers .
consider vectors @xmath396 with @xmath195 , such that @xmath450 = 2\nu_i$ ] for every @xmath451 .
assume that : 1 .
@xmath452 -12e[f_{i , n}^3 ] \to 12\nu_i^2 - 48\nu_i$ ] for every @xmath453 ; 2 .
@xmath454 whenever @xmath455 for some @xmath203 ; 3 .
@xmath456 = 0 $ ] whenever @xmath457 .
then @xmath458 where @xmath459 are independent random variables having centered @xmath432 distributions with @xmath448 degrees of freedom , respectively .
_ proof_. using the well - known carleman s condition , it is easy to check that the law of @xmath433 is determined by its moments . by corollary [ marginalblock ]
it is enough to show that condition @xmath219 of theorem [ main ] holds .
fix @xmath460 as well as @xmath461 .
switching @xmath453 and @xmath277 if necessary , assume that @xmath462 . from theorem [ noupec]@xmath219
we get that @xmath463 for each @xmath464 and every @xmath465 , except when @xmath466 . using the identity
@xmath467 ( see ( [ fubini ] ) ) together with the cauchy - schwarz inequality we infer that condition @xmath219 of theorem [ main ] holds for all values of @xmath70 , @xmath453 and @xmath277 , except of the cases : @xmath468 , @xmath469 , and @xmath470 .
assumption @xmath471 together with ( [ cov2 ] ) show that @xmath472 for all @xmath473 .
thus , it remains to verify condition @xmath219 of theorem [ main ] when @xmath470 .
lemma [ technical ] ( identity ( [ azertiop ] ) therein ) yields @xmath474 using ( [ cs ] ) and theorem [ noupec ] and a reasoning as above , it is straightforward to show that the sum @xmath475 tends to zero as @xmath21 . on the other hand , the condition on the @xmath32-th contraction in theorem [ noupec](@xmath269 ) yields that @xmath476 as @xmath21 .
moreover , we have @xmath477,\ ] ] which tends to zero by assumption @xmath405 .
all these facts together imply that + @xmath478 as @xmath21 . using ( [ cs ] ) for @xmath479
we get @xmath480 , showing that condition @xmath219 of theorem [ main ] holds true in the last remaining case .
the proof of the theorem is complete .
[ ] consider @xmath481 , where @xmath482 are even integers .
suppose that @xmath483 \to 1 , \quad e[f_{1,n}^4 ] -6e[f_{1,n}^3 ] \to -3 , \quad \text{and } \\
& e[f_{2,n}^2 ] \to 2 , \quad e[f_{2,n}^4 ] -6e[f_{2,n}^3 ] \to 0 , \quad \text{as } \ n\to \infty.\end{aligned}\ ] ] when @xmath484 or @xmath485 we require additionally : @xmath486 \to 0 \ ( q_2=2q_1).\ ] ] then theorem [ t : noupec ] ( the case @xmath487 , @xmath488 ) gives @xmath489 where @xmath490 are i.i.d .
standard exponential random variables .
[ t : bivariateblock ] let @xmath491 be positive integers .
assume further that @xmath492 .
consider @xmath493 with @xmath494 and @xmath495 .
suppose that as @xmath21 @xmath496 where @xmath497 , the marginals of @xmath498 are determined by their moments , and @xmath499 are independent . if @xmath500 \to 0 $ ] ( which trivially holds when @xmath501 ) for all @xmath248 , then @xmath502 jointly , as @xmath337 . _
proof_. we will show that condition @xmath351 of theorem [ mainblock ] holds . by we may and do assume that @xmath503=1 $ ] for all @xmath453 and @xmath19 . by theorem [ nupec]@xmath219 , @xmath504 for all @xmath505 .
observe that @xmath506 so that @xmath507 for @xmath508 , except possibly when @xmath509 . but in this latter case ,
@xmath510\big| \to 0\ ] ] by the assumption . corollary [ marginalblock ] concludes the proof .
theorem [ t : bivariateblock ] admits the following immediate corollary .
[ corindblocks ] let @xmath511 be positive integers .
consider two stochastic processes @xmath512 and @xmath513 , where @xmath514 and @xmath515 .
suppose that as @xmath21 @xmath516 where @xmath44 is centered and gaussian , the marginals of @xmath382 are determined by their moments , and @xmath517 are independent . if @xmath518 \to 0 $ ] ( which trivially holds when @xmath519 ) for all @xmath520 , then @xmath521 jointly , as @xmath337
consider a centered stationary gaussian sequence @xmath522 with unit variance . for
any @xmath523 , denote by @xmath524\ ] ] the covariance between @xmath525 and @xmath526 .
we extend @xmath70 to @xmath527 by symmetry , that is , @xmath528 . for any integer @xmath1 , we write @xmath529 to indicate the partial sums associated with the subordinated sequence @xmath530 . here
, @xmath531 denotes the @xmath5th hermite polynomial given by ( [ hermitepol ] ) .
the following result is a summary of the main finding in breuer and major @xcite .
[ bm ] if @xmath532 then , as @xmath21 , @xmath533 where @xmath6 is a standard brownian motion and @xmath534^{1/2}$ ] .
assume further that the covariance function @xmath70 has the form @xmath535 with @xmath536 and @xmath537 a function which is slowly varying at infinity and bounded away from @xmath538 and infinity on every compact subset of @xmath539 .
the following result is due to taqqu @xcite .
[ taqqu ] if @xmath540 then , as @xmath21 , @xmath541 where @xmath542^{-1/2}$ ] and @xmath543 is a rosenblatt process of parameter @xmath544 , defined as @xmath545 with @xmath546 @xmath547 an explicit constant such that @xmath548=1 $ ] , and the double wiener - it integral @xmath549 is with respect to a two - sided brownian motion @xmath6 .
let @xmath550 be an integer .
the following result is a consequence of corollary [ corindblocks ] and theorems [ bm ] and [ taqqu ] .
it gives the asymptotic behavior ( after proper renormalization of each coordinate ) of the pair @xmath551 when @xmath552 .
since what follows is just mean to be an illustration , we will not consider the remaining case , that is , when @xmath553 ; it is an interesting problem , but to answer it would be out of the scope of the present paper .
[ bm - dmt ] let @xmath550 be an integer , and let the constants @xmath554 and @xmath555 be given by theorems [ bm ] and [ taqqu ] , respectively . 1 . if @xmath556 then @xmath557 where @xmath558 is a standard brownian motion in @xmath559 .
2 . if @xmath560 then @xmath561 where @xmath6 is a brownian motion independent of the rosenblatt process @xmath562 of parameter @xmath563 .
_ proof _ : let us first introduce a specific realization of the sequence @xmath522 that will allow one to use the results of this paper .
the space @xmath564 being a real separable hilbert space , it is isometrically isomorphic to either @xmath565 ( for some finite @xmath566 ) or @xmath567 .
let us assume that @xmath568 , the case where @xmath569 being easier to handle .
let @xmath570 be an isometry .
set @xmath571 for each @xmath572 .
we have @xmath573=\int_0^\infty e_k(x)e_l(x)dx,\quad k , l{\geqslant}1.\label{ekrho}\ ] ] if @xmath574 denotes a standard brownian motion , we deduce that @xmath575 these two sequences being indeed centered , gaussian and having the same covariance structure . using ( [ mapping ] ) we deduce that @xmath576 has the same distribution than @xmath577 ( with @xmath578 the @xmath5-tuple wiener - it integral associated to @xmath6 ) . hence , to reach the conclusion of point 1 it suffices to combine corollary [ corindblocks ] with theorem [ bm ] . for point 2 , just use corollary [ corindblocks ] and theorem [ taqqu ] , together with the fact that the distribution of @xmath579 is determined by its moments ( as is the case for any double wiener - it integral ) . to develop the next application we will need the following basic ingredients : 1 . a sequence @xmath580 of i.i.d . random variables , with mean 0 , variance 1 and all moments finite .
two positive integers @xmath581 as well as two sequences @xmath582 , @xmath239 of real - valued functions satisfying for all @xmath583 and @xmath584 , 1 .
@xmath585 for every permutation @xmath117 ; 2 .
@xmath586 whenever @xmath587 for some @xmath588 ; 3 . @xmath589 .
consider @xmath590 this series converges in @xmath373 , @xmath591=0 $ ] and @xmath592=1 $ ] .
we have the following result . [ thmmoo ] as @xmath21 , assume that the contribution of each @xmath231 to @xmath593 is uniformly negligible , that is , @xmath594 and that , for any @xmath209 , @xmath595 then @xmath596 and @xmath597 are asymptotically moment - independent . _ proof : _ fix @xmath598 .
we want to prove that , as @xmath21 , @xmath599-e[q_{1,n}({\bf x})^m]e[q_{2,n}({\bf x})^n]\to 0.\ ] ] the proof is divided into three steps . _ step 1_. in this step we show that @xmath600-e[q_{1,n}({\bf g})^mq_{2,n}({\bf g})^n]\to 0 \quad \text{as } \
n \to \infty.\ ] ] following the approach of mossel , odonnel and oleszkiewicz @xcite , we will use the lindeberg replacement trick .
let @xmath601 be a sequence of i.i.d .
@xmath602 random variables independent of @xmath603 . for a positive integer @xmath604 , set @xmath605 , and put @xmath606 .
fix @xmath607 and write for @xmath584 and @xmath104 , @xmath608 where @xmath609 means that the term @xmath610 is dropped ( observe that this notation bears no ambiguity : indeed , since @xmath611 vanishes on diagonals , each string @xmath612 contributing to the definition of @xmath613 contains the symbol @xmath604 exactly once ) . for each @xmath604 and @xmath348 , note that @xmath614 and @xmath613 are independent of the variables @xmath615 and @xmath616 , and that @xmath617 by the binomial formula , using the independence of @xmath615 from @xmath614 and @xmath613 , we have @xmath618 \\ & = \sum_{i=0}^m\sum_{j=0}^n \binom{m}{i}\binom{n}{j}e[u_{1,n , s}^{m - i}u_{2,n , s}^{n - j}v_{1,n , s}^{i}v_{2,n , s}^j]e[x_s^{i+j}].\end{aligned}\ ] ] similarly , @xmath619 \\ &
= \sum_{i=0}^m\sum_{j=0}^n \binom{m}{i}\binom{n}{j}e[u_{1,n , s}^{m - i}u_{2,n , s}^{n - j}v_{1,n , s}^{i}v_{2,n , s}^j]e[g_s^{i+j}].\end{aligned}\ ] ] therefore @xmath618 - e[q_{1,n}({\bf w}^{(s)})^m q_{2,n}({\bf w}^{(s)})^n ] \\ & = \sum_{i+j \ge 3 } \binom{m}{i}\binom{n}{j } e[u_{1,n , s}^{m - i}u_{2,n , s}^{n - j}v_{1,n , s}^{i}v_{2,n , s}^j]\big ( e[x_s^{i+j } ] - e[g_s^{i+j}]\big).\end{aligned}\ ] ] now , observe that propositions 3.11 , 3.12 and 3.16 of @xcite imply that both @xmath620 and @xmath621 are uniformly bounded in all @xmath622 spaces .
it also implies that , for any @xmath623 , @xmath584 and @xmath624 , @xmath625^{1/p } \leq c_p \ , e[v_{k , n , s}^2]^{1/2},\ ] ] where @xmath626 depends only on @xmath627 .
hence , for @xmath628 , @xmath629 , @xmath630 , we have @xmath631\big| \le c\ , e[v_{1,n , s}^2]^{i/2 } \ , e[v_{2,n , s}^2]^{j/2},\ ] ] where @xmath632 does not depend on @xmath633 .
since @xmath634=e[g_i]=0 $ ] and @xmath635=e[g_i^2]=1 $ ] , we get @xmath636 = q_kq_k!\sum_{i_2,\ldots , i_{q_k}=1}^{\infty } a_{k , n}(s , i_2,\ldots , i_{q_k})^2.\ ] ] when @xmath637 , then is bounded from above by @xmath638 where @xmath632 does not depend on @xmath633 , and we get a similar bound when @xmath639 .
if @xmath640 , then @xmath641 ( @xmath642 ) , so isz bounded from above by @xmath643 and we have a similar bound when @xmath644 .
taking into account assumption we infer that the upper - bound for is of the form @xmath645 where @xmath646 and @xmath632 is independent of @xmath647 .
we conclude that @xmath648 - e[q_{1,n}({\bf w}^{(s)})^m q_{2,n}({\bf w}^{(s)})^n ] \big| \\ & & \hskip6 cm \le c \epsilon_n \sum_{k=1}^2 \sum_{i_2,\ldots , i_{q_k}=1}^{\infty } a_{k , n}(s , i_2,\ldots , i_{q_k})^2,\end{aligned}\ ] ] where @xmath632 does not depend on @xmath647 . since , for fixed @xmath649 , @xmath650 in @xmath373 as @xmath651 , by propositions 3.11 , 3.12 and 3.16 of @xcite , the convergence holds in all @xmath622 . hence @xmath652-e[q_{1,n}({\bf g})^mq_{2,n}({\bf g})^n]\big| \\ & \le \sum_{s=1}^{\infty } \big| e[q_{1,n}({\bf w}^{(s-1)})^m q_{2,n}({\bf w}^{(s-1)})^n ] - e[q_{1,n}({\bf w}^{(s)})^m q_{2,n}({\bf w}^{(s)})^n ] \big| \\ & \le c \epsilon_n \sum_{k=1}^2 \sum_{i_1,\ldots , i_{q_k}=1}^{\infty } a_{k , n}(i_1,i_2,\ldots , i_{q_k})^2 = c \big((q_1!)^{-1}+ ( q_2!)^{-1}\big ) \epsilon_n . \end{aligned}\ ] ] this proves .
_ step 2_. we show that @xmath337 , @xmath653-e[q_{1,n}({\bf g})^m]\to 0\quad\mbox{and}\quad e[q_{2,n}({\bf x})^n]-e[q_{2,n}({\bf g})^n]\to 0.\ ] ] the proof is similar to step 1 ( and easier ) .
thus , we omit it . _ step 3_. without loss of generality we may and do assume that @xmath654 , where @xmath6 is a standard brownian motion . for @xmath584 and @xmath239 , due to the multiplication formula ( [ multiplication ] ) , @xmath655 is a multiple wiener - it integral of order @xmath656 with respect to @xmath6 : @xmath657 \times\ldots\times [ i_{q_k}-1,i_{q_k } ] } \right).\ ] ] in this setting , condition ( [ mixed ] ) coincides with condition ( @xmath207 ) of theorem [ main ] ( or ( @xmath658 ) of theorem [ mainblock ] ) . therefore , @xmath659-e[q_{1,n}({\bf
g})^m]e[q_{2,n}({\bf g})^n]\to 0.\ ] ] combining ( [ moocsq ] ) , ( [ moocsq2 ] ) and ( [ mixedcsq ] ) we get the desired conclusion ( [ goal ] )
. the conclusion of theorem [ thmmoo ] may fail if either ( [ influence ] ) or ( [ mixed ] ) are not satisfied .
it follows from step 3 above that the theorem fails when does not hold and @xmath603 is gaussian .
theorem [ thmmoo ] also fails when ( [ influence ] ) is not satisfied , ( [ mixed ] ) holds , and @xmath603 is a rademacher sequence , as we can see from the following counterexample .
consider @xmath660 , and set @xmath661 then @xmath662 and @xmath663 , where @xmath231 are i.i.d .
with @xmath664 .
it is straightforward to check that ( [ mixed ] ) holds and obviously ( [ influence ] ) is not satisfied . since @xmath665
, we get @xmath666\neq e[q_{1,n}({\bf x})^2]e[q_{2,n}({\bf x})^2],\ ] ] implying in particular that @xmath596 and @xmath597 are ( asymptotically ) moment - dependent .
we are grateful to jean bertoin for useful discussions and to ren schilling for reference @xcite .
we warmly thank an anonymous referee for suggesting a shorter proof of theorem [ main ] ( which evolved into the proof of a more general statement , theorem [ mainblock ] ) and for useful comments and suggestions , which together with the editor s constructive remarks , have led to a significant improvement of this paper .
g. peccati and c.a .
tudor ( 2005 ) .
gaussian limits for vector - valued multiple stochastic integrals . in : _
sminaire de probabilits xxxviii , 247 - 262 .
lecture notes in math . *
1857 * , springer - verlag , berlin .
_ j. rosiski and g. samorodnitsky ( 1999 ) .
product formula , tails and independence of multiple stable integrals .
_ advances in stochastic inequalities _
( atlanta , ga , 1997 ) , 169 - 194 , _ contemp . math .
_ * 234 * , amer .
soc . , providence , ri . | we characterize the asymptotic independence between blocks consisting of multiple wiener - it integrals . as a consequence of this characterization ,
we derive the celebrated fourth moment theorem of nualart and peccati , its multidimensional extension , and other related results on the multivariate convergence of multiple wiener - it integrals , that involve gaussian and non gaussian limits .
we give applications to the study of the asymptotic behavior of functions of short and long range dependent stationary gaussian time series and establish the asymptotic independence for discrete non - gaussian chaoses . |
we appreciate dr .
stphane mangin of nancy universit and dr .
eric e. fullerton of the university of california , san diego for providing the spin - valve samples used in the characterization of magnetic properties used in this study .
we also acknowledge marius klein for support in the initial stages of developing these measurements .
this research was supported at nyu by nsf grant dmr-1006575 . | we report on a technique for determining electronic transport properties using commercially available audio cards . using a typical 24-bit audio card simultaneously as a sine wave generator and a narrow bandwidth ac voltmeter
, we show the spectral purity of the analog - to - digital and digital - to - analog conversion stages , including an effective number of bits greater than 16 and dynamic range better than 110 db .
we present two circuits for transport studies using audio cards : a basic circuit using the analog input to sense the voltage generated across a device due to the signal generated simultaneously by the analog output ; and a digitally - compensated bridge to compensate for nonlinear behavior of low impedance devices .
the basic circuit also functions as a high performance digital lock - in amplifier .
we demonstrate the application of an audio card for studying the transport properties of spin - valve nanopillars , a two - terminal device that exhibits giant magnetoresistance ( gmr ) and whose nominal impedance can be switched between two levels by applied magnetic fields and by currents applied by the audio card
. studies of the electronic transport properties of devices and materials require specialized instrumentation .
most transport studies break down into two classes : current - voltage characterization and differential resistance measurements .
the former can be done using a sourcemeter , which supplies power to a device and simultaneously measures both the voltage drop across the device and current flowing through the device .
differential resistance is typically measured using a lock - in amplifier
. a lock - in amplifier is a robust solution for accurate , low - noise measurements of the voltage response @xmath0 to a small excitation current , @xmath1 .
this type of measurement has a narrow noise bandwidth and can be conducted above the 1/@xmath2 noise threshold with high dynamic reserve .
however , it is sometimes desirable to measure the static response ( current - voltage characteristics ) simultaneously with the differential response , @xmath0/@xmath1 .
this can not be done with only a lock - in amplifier and requires the addition of an additional source - measure instrument .
furthermore , commercially available scientific instruments for both measurement tasks can cost upwards of $ 10,000 .
commercially available audio cards can output and record arbitrary waverforms and do many of the tasks typically reserved for source - measure units and lock - in amplifiers .
these are low - cost , versatile instruments , with 24-bit processing and functionality as a signal generator with less than -100 dbc harmonic distortion .
audio cards also can function as a narrow bandwidth ac voltmeter , with better than 110 db dynamic range . with such high resolution
, these devices outpace many direct digital synthesizers as a pure sine - wave generator and rank with commercial lock - in amplifiers in dynamic rejection .
furthermore , audio cards typically have multiple synchronized input - outputs which permit simultaneous characterization of multiple devices .
we present the application of an usb - based audio card - the e - mu 0404 ( creative professional ) - for static and dynamic transport measurements , which can be assembled very economically compared to commercial scientific instruments @xcite .
this measurement device permits sourcing and detecting differential signals @xmath0/@xmath1 as well as sensing current - voltage characteristics simultaneously .
we will present measurements of magnetic nanostructures , whose current - voltage and differential resistance are used to determine their underlying magnetic and electronic properties .
our measurement setup is centered around a single e - mu 0404 connected via usb to a pc .
the audio card possesses two balanced high - impedance ( 1 m@xmath3 ) inputs and two unbalanced analog outputs ( @xmath4 = 2 v ) .
analog - to - digital ( a / d ) and digital - to - analog ( d / a ) conversion is done with a variable sampling rate ( 44.1 , 48 , 88.2 , 96 , 176.4 and 192 khz from an internal crystal ) with 24-bit input / output ( i / o ) processing . also included are optical digital interconnects ( s / pdif ) for synchronous i / o over multiple audio cards .
the data is read from the audio card using the steinberg audio stream input / output ( asio sdk ) drivers , which are controlled using a python wrapper ( pyaudio ) around portaudio c programming libraries @xcite .
we have tested some standard spectral characteristics of the audio card , many of which are described in electronics textbooks @xcite .
figure [ fig : nsd ] demonstrates the noise spectral density of the a / d converter at the input stage . for a variety of input resistances , the noise density generated from a fast fourier transform ( fft ) of the input signal ( bandwidth @xmath5 = 0.2 hz ) compared with the johnson noise is nearly flat across the frequency spectrum ( 20 hz-20 khz ) .
the noise amplitude gets closer to the johnson noise for larger impedances , which could suggest that voltage noise at the input stage plays a role for the smaller load impedances .
furthermore , the input noise is only one order of magnitude larger than the best input noise for a commercial lock - in amplifier @xcite .
figure [ fig : sinad ] illustrates the spectral purity of the entire audio card by looping the output of the d / a converter ( a sine wave at 6 db below full scale ) back into the a / d converter without preamplification
. the calculated ratio of signal to noise and distortion ( sinad ) from an fft of the input signal leads directly to the effective number of bits ( enob ) , which is a measure of the quality of the i / o digitization . a nearly constant value of 16 bits indicates that of the full 24 bits resolution , the upper 16 bits contain useful information above the noise floor , which is critical for sensing small signal changes over a larger background signal .
other useful spectral properties of the measured output include total harmonic distortion plus noise ( thdpn ) ( -99.5 dbc ( decibels below carrier ) ) , signal - to - noise ( -100.5 dbc ) , spurious free dynamic range ( -112 dbc ) , and the noise floor ( -132 dbc ) .
thus , we have a i / o device that functions as an ultra - low distortion , 24-bit resolution signal generator with an a / d input stage that measures low and higher - impedance devices alike with relatively flat noise density over the range of audible frequencies . figure [ fig : schematic ] illustrates the configuration we use for transport measurements .
the top - left block represents the usb - based audio card , whose two outputs , l and r , are amplified and fed into two 1 @xmath6 resistors ( chosen to be two orders of magnitude larger than the sample resistance ) to create two stiff current sources , which feed into the sample ( @xmath7 ) and a potentiometer ( @xmath8 ) .
leads across the terminals of @xmath7 feed into the `` @xmath9 '' audio card input and permit direct measurement of the sample i - v characteristic and/or differential resistance .
we also implement a digitally compensated bridge measurement using the audio card .
balancing bridge designs based upon active compensation have been well established for systems where the versatility to compensate for arbitrary reactances in complex networks where capacitances and inductances need to be compensated , or even for systems where small nonlinear deviations in i - v characteristics require versatile and precise balancing @xcite . by outputing an appropriate signal on channel l , any parasitics or reactances of `` @xmath7 '' can be compensated for .
this circuit permits us to compensate nonlinearities of the device i - v characteristic as well as unaccounted series reactive elements .
it also functions as an inexpensive impedance meter across a wide frequency band ( 20 hz - 20 khz ) .
audio card measurements are especially useful for fast characterization of spin - valve nanopillars - a two terminal magnetic device composed of two uniaxial ferromagnetic layers that exhibits two stable resistance states depending on the relative magnetization orientation of the layers ( see the inset of figure [ fig : rvsh ] ) .
the spin - valve state can be toggled between high ( anti - parallel ) and low resistance ( parallel ) by applied magnetic fields ( field - induced magnetization switching ) or by electrical currents ( current - induced magnetization switching ) .
we can determine the relative orientation of the layers by measuring the device resistance ( static or differential ) as a function of the applied magnetic field .
however , the resistance change is small in comparison to the device resistance and the joule heating - related nonlinearity , which requires that we measure small changes in resistance and differential resistance in order to observe the device switch between high and low resistance states .
measuring the differential resistance versus field of a spin - valve nanopillar allows us to demonstrate the use of the audio card as a lock - in amplifier .
we used the audio card to source a 882 hz excitation current ( 200 @xmath10a rms amplitude ) to our device , while simultaneously using the audio card to measure the generated voltage .
we then multiplied the digitized voltage by a unit amplitude sine and cosine of the same frequency to get the in - phase and out - of - phase components of our signal in relation to our reference and use a digital , first - order butterworth lowpass filter ( time constant 30 ms ) to eliminate noise and the @xmath11 component , and then calculate the magnitude from the in - phase and out - of - phase components , plotting @xmath0/@xmath1 as a function of applied magnetic field in figure [ fig : rvsh ] @xcite .
sharp changes in the differential resistance indicate toggling between parallel and antiparallel magnetic orientations of the two ferromagnetic layers .
there is an interplay between applied magnetic fields and electric currents in the orientation of the two magnetic layers , which has been examined previously and is quite important for applications like magnetic random access memory @xcite .
this relationship is usually plotted in the form of a state diagram , which maps out under a given applied magnetic field and direct current the boundary of different microscopic magnetization configurations of the two spin - valve layers as determined by sharp changes in measured giant magnetoresistance ( gmr ) .
thus far , studies of the spin - valve state diagram have been limited by the speed of the measurement , which can take several days to complete in order to explore the wide parameter space of applied magnetic fields and electric currents .
we present a method for quickly mapping out the field - current parameter space of a spin - valve nanopillar with the use of the audio card in the direct measurement ( i - v ) configuration . in figure [
fig : magnetic ] , our experimental results are displayed .
we measure the voltage generated across the sample , @xmath7 , on channel `` @xmath9 '' , under a linear current ramp in the form of a 441 hz triangle wave with peak ampllitude @xmath12 = 20 ma flowing through a spin - valve nanopillar device whose cross sectional area is @xmath13nm@xmath14 .
the physical properties of this device have been described in detail elsewhere @xcite . in fig .
[ fig : magnetic](a ) the voltage across the nanopillar is measured by the analog - to - digital converter ( d / a ) simultaneous to the application of the triangle wave from the audio card s digital - to - analog ( d / a ) converter ) . figure [ fig : magnetic](b ) demonstrates the corresponding gmr effect for current switching of the spin - valve from an antiparallel ( high - resistance ) to parallel ( low - resistance ) configuration and back again . in order to map out the state diagram
, we generate a triangle field ramp from our electromagnet with period t = 8 s and @xmath15 = 200 mt on top of our continuous ac current . over one half - period of the field cycle we can generate several hundred gmr curves like the one shown in fig . [
fig : magnetic](b ) , and by monitoring the currents and fields at which the device resistance toggles between high and low , we arrive at the state diagram in fig .
[ fig : magnetic](c ) .
this diagram illustrates the outer regions in which the spin - valve device is always oriented p or ap , and where the device exhibits bistability or hysteresis in the inner regions of the parameter ( h , i ) space . in summary ,
we have demonstrated the use of an inexpensive , commercially - available audio card for high quality measurements of nanostructures .
we believe that this can save researchers a significant amount of money on otherwise expensive standard scientific measurement tools while affording greater flexibility for signal processing , multiple channels for parallel studies , high dynamic range and most importantly , high speed capabilities .
we have shown the spectral properties that make the e - mu 0404 usb audio card competitive with commercial laboratory instruments .
the audio card unit is versatile - it can function as a tool for direct measurement of devices or it can be integrated into a digitally compensating bridge .
while we have demonstrated the use of the audio card as a source - measure unit and as a digital lock - in amplifier for the study of low - impedance spin - valve devices , it should be noted that this measurement technique can be used for many types of measurements where a high quality signal source is required . |
the monitored drift tube ( mdt ) chambers of the atlas muon spectrometer @xcite have demonstrated that they provide very precise and robust tracking over large areas .
goals of atlas muon detector upgrades are to increase the acceptance for triggering ande precise muon momentum measurement and to improve the rate capability of the muon chambers in the high - background regions at increasing lhc luminosity .
small - diameter muon drift tube ( smdt ) chambers have been developed for these purposes .
with half the drift - tube diameter of the mdt chambers ( 15 instead of 30 mm ) and otherwise unchanged operating parameters , including a gas gain of 20000 in ar : co@xmath1 ( 93:7 ) gas mixture at 3 bar , smdt chambers share the advantages of the mdts , but have an order of magnitude higher rate capability @xcite and can be installed in detector regions where mdt chambers do not fit in .
the construction of twelve smdt chambers to be installed in the feet of the atlas detector in the middle layer of the barrel muon spectrometer in the lhc winter shutdown 2016/17 is in progress since october 2014 .
they consist of 356 drift tubes of 1120 mm length in two quadruple layers ( multilayers ) which are separated by an aluminum spacer and support frame .
half of the chambers has already been constructed .
the purpose of this upgrade is to increase the acceptance for the three - point measurement of muon tracks which will result in an substantial improvement of the muon momentum resolution .
two similar chambers with 624 tubes of 2150 mm length have already been installed in 2014 in the middle layer of the bottom barrel sector of the spectrometer for the same purpose .
for operation at hl - lhc , replacement of the mdt chambers in the barrel inner layer by smdt chambers with integrated thin - gap rpc chambers is foreseen in order to increase the muon trigger acceptance and robustness as well as the rate capability of the precision tracking chambers . at the same time the drift distance information from the mdt and smdt chambers
will be included in the first - level muon trigger . in the lhc
shutdown 2019/20 , 16 such smdt and rpc chamber packages will already be installed at the ends of the barrel inner layer for lhc run 3 .
the chamber design and construction procedures have been optimized for mass production while providing highest mechanical accuracy in the sense wire positioning .
the smdt drift tubes are standard industrial aluminium tubes with 15 mm diameter and a wall thickness of 0.4 mm , chromatised for cleaning and reliable electrical ground contact .
first , the drift tubes are assembled using a semi - automated wiring machine in a temperature - controlled clean room .
the sense wires are fed through the tubes and the endplugs ( see fig . [ fig::endplug ] ) without manual contact by means of air flow and fixed in copper crimp tubelets after tensioning to @xmath2 g corresponding to a gravitational sag of only @xmath3 m ( absolute tolerances ) .
the wires are positioned at each tube end with few micron precision with respect to a cylindrical external reference surface on the central brass insert of the endplug which also holds the spiral shaped wire locator on the inside of the tube .
the drift tubes are sealed with the endplugs via two o - rings each using mechanical swaging of the tube walls . for the injection
molded endplug insulators and gas connectors for the individual tubes , plastic materials with minimum outgassing have been selected which also are immune against cracking .
no aging has been observed in irradiation tests of the smdt drift tubes up to 9 c / cm charge accumulated on the wire .
the 4300 drift tubes of the twelve new chambers have been produced at a typical rate of 100 tubes per day and a failure rate of less than @xmath4 due to intermittent failures of the wire tensioning and crimping tools detected by visual inspection during drift tube assembly .
the critical performance parameters , wire tension , gas leak rate ( @xmath5 bar l / s ) and leakage current ( @xmath6 na / m ) at nominal operating voltage ( 2730 v ) are measured for every tube before assembly in a chamber .
the overall failure rate of these tests was only @xmath7 .
the drift tubes are assembled into chambers in a climatised clean room by inserting the endplug reference surfaces into a grid of fitting bores in the assembly jigs at each chamber end which define the wire positions with an accuracy of better than 5 micron and glueing them together and to the spacer and support frame using an automated glue dispenser .
a complete chamber can be assembled within one working day .
the precise mounting of the platforms for the optical alignment sensors with respect to the wires requires an additional day . before the installation of the parallel gas distribution system ( which fulfills the stringent leak rate requirement of less than 0.2 mbar / h ) and of the specially designed electronics boards with 6 x 4 readout channels matching the transverse cross section of the tube multilayers , the positions of the individual sense wires are measured at the two chamber ends via the endplug reference surfaces with an automated coordinate measuring machine with a precision of about 2 @xmath0 m ( rms ) . at the same time , the relative angle between the wire grids at the two chamber ends ( chamber torsion ) and the positions and orientations of the optical alignment sensor mounting platforms are measured .
sense wire positioning accuracies of better than 10 @xmath0 m ( rms ) are routinely achieved , surpassing the requirement for the atlas mdt chambers of 20 @xmath0 m ( rms ) and allowing for a spatial resolution of a chamber with 2 x 4 tube layers of better than 35 @xmath0 m . in the best cases , the ultimate wire positioning accuracy of 5 @xmath0 m ( rms ) as given by the precision of the assembly jigs
has been reached ( see fig . [ fig::residuals ] ) .
after the measurement , the wire positions are known with 2 @xmath0 m accuracy . | atlas muon detector upgrades aim for increased acceptance for muon triggering and precision tracking and for improved rate capability of the muon chambers in the high - background regions of the detector with increasing lhc luminosity .
the small - diameter muon drift tube ( smdt ) chambers have been developed for these purposes . with
half of the drift - tube diameter of the mdt chambers and otherwise unchanged operating parameters , smdt chambers share the advantages of the mdts , but have an order of magnitude higher rate capability and can be installed in detector regions where mdt chambers do not fit in .
the chamber assembly methods have been optimized for mass production , minimizing construction time and personnel .
sense wire positioning accuracies of 5 @xmath0 m have been achieved in serial production for large - size chambers comprising several hundred drift tubes .
the construction of new smdt chambers for installation in the 2016/17 winter shutdown of the lhc and the design of smdt chambers in combination with new rpc trigger chambers for replacement of the inner layer of the barrel muon spectrometer are in progress . |
by _ a tensor category over a commutative ring _ @xmath3 we understand a @xmath3-linear - modules and the compositions are @xmath3-bilinear , for instance , if @xmath4 this means that the category is preadditive . ]
symmetric monoidal category @xmath5 where @xmath6 is the tensor product bifunctor , @xmath7 is the associativity transformation , @xmath8 is the symmetric braiding , @xmath9 is the monoidal identity , @xmath10 and @xmath11 are left and right unity transformations .
we require the structures to agree : the tensor product must be @xmath3-bilinear on morphisms and the natural morphism @xmath12 must be a ring isomorphism .
we attempt to follow notations and terminology of joyal and street @xcite , although their `` tensor category '' is the same as a monoidal category , while our tensor categories are @xmath3-linear and symmetric .
observe that @xmath13 is not required to be abelian or even additive .
additivity can be easily fixed by completing the category with respect to finite direct sums and the zero object but not being abelian is an important distinction with other `` tensor categories '' that appear in the literature .
our tensor categories are close to those of hinich and vaintrob @xcite with an exception of our additional assumption that @xmath14 is an isomorphism .
this assumption is not very restrictive .
suppose @xmath15 is a symmetric monoidal additive category
. then the ring @xmath16 is commutative and the category @xmath13 is a tensor category of @xmath17 .
the ring @xmath17 is commutative by the eckmann - hilton argument ( * ? ? ?
* prop 6.2 ) .
each @xmath18 is a @xmath17-@xmath17-bimodule via the left action @xmath19 and the right action @xmath20 these actions coincide : given @xmath21 , @xmath22 , both @xmath23 and @xmath24 can be read off the diagram @xmath25 as compositions from the top left corner to the bottom right corner .
the diagram is commutative : the outside squares are commutative by the standard fact ( * ? ? ?
* prop 2.1 ) while the middle square is commutative because the commutativity constraint @xmath26 is a natural transformation .
hence , @xmath27 .
finally , @xmath17-bilinearity of compositions is evident .
tensor categories are ubiquitous in modern mathematics , so we spare the reader of an example now but introduce some as we go along . given an object @xmath28 in a tensor category over @xmath3 , we define its iterated tensor products by @xmath29 and @xmath30 . there is an action of the symmetric group @xmath31 on the object @xmath32 , i.e. a semigroup homomorphism @xmath33 defined as follows . using a chain of associativity constraints @xmath34 we define it on transpositions by @xmath35 and extend to the whole group : existence of such an extension follows from the axioms of symmetric monoidal category .
it gives a @xmath36-module structure on @xmath37 for a pair of objects @xmath38 of @xmath13 .
given a tensor category @xmath13 over @xmath3 and ring homomorphisms @xmath39 and @xmath40 of commutative rings , we can construct a new tensor category @xmath41 over @xmath42 .
it has the same objects as @xmath13 but new morphisms @xmath43 .
the compositions are defined in the obvious way : @xmath44 .
similarly , all the natural transformations are extended to @xmath45 .
all verifications are routine and left to an interested reader .
let @xmath13 and @xmath46 be tensor categories over the same commutative ring @xmath3 .
_ a tensor functor _
@xmath47 consists of a @xmath3-linear functor @xmath48 , a natural isomorphism @xmath49 and an isomorphism @xmath50 .
they have to satisfy the standard axioms @xcite .
we can talk about algebras in @xmath13 over a @xmath3-linear operad or even a @xmath3-linear prop @xcite .
_ a @xmath3-linear prop _ is a tensor category @xmath51 over @xmath3 satisfying the following five conditions : its objects form the monoid @xmath52 , i.e. @xmath53 , it is strict , in this context it means that all @xmath54 , @xmath55 , and @xmath56 , for each object @xmath57 a semigroup embedding @xmath58 , @xmath59 is chosen , for all @xmath60 , @xmath61 their tensor product is another permutation @xmath62 : @xmath63 the commutativity constraint is the shuffle @xmath64 : @xmath65 now _ an algebra in _ @xmath13 over a prop @xmath51 is a tensor functor @xmath66 .
this level of generality is not always useful as the prop needs to be constructed .
a more elementary approach is more suitable for our purposes . _
a signature _ @xmath67 is a set of operations @xmath68 with head and tail functions @xmath69 , so that we think of @xmath70 as an operation on @xmath71 variables with @xmath72 values . _
an algebra of signature _
@xmath68 is a triple @xmath73 where @xmath13 is a tensor category over a commutative ring @xmath3 , @xmath28 is an object of @xmath13 , and @xmath74 is a family of morphisms . if @xmath3 or @xmath13 is fixed , we talk about @xmath3-algebras or @xmath13-algebras .
we specify a signature by listing the size of @xmath68 and pairs @xmath75 , thinking that @xmath76 and the @xmath77-th pair are the tail and head values of @xmath78 .
let us contemplate the connection between these two notions of algebra .
a prop @xmath51 admits a generating set of morphisms @xmath79 .
an algebra @xmath80 over a prop @xmath51 determines the algebra @xmath81 of the signature @xmath67 , determined by the generating set . in the opposite direction ,
an algebra @xmath73 of the signature @xmath67 may or may not determine an algebra over @xmath51 . in general , it determines only an algebra over a free prop @xmath82 .
the prop @xmath51 is a quotient of the free prop @xmath82 , so to determine an algebra over @xmath51 the algebra must satisfy some axioms , coming from @xmath51 .
thus , the props provide an implicit method of requesting axioms on algebras , but often it is more convenient to do explicitly . in the partial case of an algebra @xmath28 of signature
@xmath83 $ ] the axioms are just polylinear identities .
let @xmath84 be the free nonassociative algebra in countably many variables , @xmath85 the subset of polylinear elements of degree @xmath86 , linear in fixed variables , say @xmath87 .
there is a natural `` evaluation '' map of @xmath3-modules @xmath88 defined recursively on monomials : @xmath89 where @xmath90 and @xmath91 is the composition of associativity constraints , and then extended to linear combinations by @xmath92 .
now we say that @xmath28 satisfies a polylinear identity @xmath93 if @xmath94 .
_ a lie algebra _ is an algebra of signature @xmath83 $ ] satisfying the jacobi and the anticommutativity identities : @xmath95 which can be written explicitly in the tensor notation as @xmath96 notice that if @xmath13 is the category of @xmath3-modules and @xmath97 then our definition is just the usual definition of a lie algebra over @xmath3 . however , if @xmath98 , our anticommutativity is weak , i.e. , @xmath99 but we can not conclude that @xmath100 .
_ an associative algebra _ is an algebra of signature @xmath101 $ ] satisfying the associativity @xmath102 and the left and right unities .
the latter can not be written in terms of @xmath103 so we recourse to the tensor notation formulating the axioms : @xmath104 _ a metric object _ is an algebra of signature @xmath105 $ ] such that both structures are symmetric : @xmath106 and @xmath28 is a dual object of @xmath28 in the sense of monoidal categories , i.e. , the identity @xmath107 holds .
notice that @xmath108 is `` non - degenerate '' in this case .
if @xmath109 is another dual object , then @xmath110 is a canonical isomorphism .
following hinich and vaintrob @xcite , we are introducing casimir and metric lie algebras in the next section .
let @xmath111 be a lie algebra . _
a representation _ of @xmath112 is a pair @xmath113 where @xmath114 is an object in @xmath13 , @xmath115 and the jacobi identity @xmath116 holds .
completely parallel to the case of lie algebras in vector spaces , the representations form a tensor category @xmath117 .
the morphisms @xmath118 are those morphisms in @xmath119 that commute with the action of @xmath112 .
the tensor product is the tensor product in @xmath13 with the usual action : @xmath120 the constraints are inherited from @xmath13 .
let us summarise this in a proposition whose proof is left to an interested reader .
let @xmath112 be a lie @xmath13-algebra .
then the category @xmath117 is a tensor category over @xmath17 .
its identity object is @xmath121 with zero action .
the multiplication in @xmath112 is a morphism in @xmath117 , so @xmath112 is a lie @xmath117-algebra .
one could ask what sort of tensor categories one gets by iteration : consider representations of @xmath112 in @xmath117 , etc .
the answer is rather uninteresting : the reader can easily verify that subject to the existence of finite direct sums in @xmath13 , the category of representations of @xmath112 in @xmath117 is canonically equivalent to the category of representations of the semidirect product @xmath122 in @xmath13 ( where @xmath123 is the adjoint representation of @xmath112 treated as a lie algebra with the zero lie product ) .
now _ a casimir lie algebra _ is a @xmath13-algebra of signature @xmath124 $ ] such that @xmath125 defines a lie algebra while @xmath108 is a symmetric morphism in the category of @xmath112-modules .
similarly _ a metric lie algebra _ is a @xmath13-algebra @xmath112 of signature @xmath126 $ ] such that @xmath127 is a lie @xmath13-algebra , and @xmath128 is a metric @xmath117-object .
naturally there are two notions of a homomorphism between algebras of the same signature . _ a homomorphism from _
@xmath129 _ to _
@xmath130 is a morphism @xmath131 which commutes with all multiplications , i.e. the diagram @xmath132 is commutative for all @xmath133 . to avoid confusion we use the term specialization for the second notion .
_ a specialization from _ @xmath129 _ to _
@xmath134 is a triple @xmath135 where @xmath136 is a ring homomorphism , @xmath137 is a tensor functor , and @xmath138 is a homomorphism from @xmath139 to @xmath140 .
a specialization such that @xmath138 is an isomorphism is called _
isospecialization_. notice that we may consider the three types of homomorphisms ( or specializations ) between metric lie algebras : a lie algebra homomorphism , a casimir lie algebra homomorphism , and a metric lie algebra homomorphism .
the difference is in the preserved operations .
varieties of algebras is a topic of an active study in algebra @xcite .
should one study varieties of algebras in categories ? in this subsection
we make some easy observations and pose some questions . at the first glance ,
at least if @xmath3 is a @xmath141-algebra , the varieties of algebras in categories are not much different from the varieties of usual algebras .
indeed , an interested reader could easily verify that the identities of a @xmath13-algebra is equal to the set of polylinear elements in some @xmath142-ideal ( an ideal closed under substitutions ) of the free algebra @xmath143 .
for instance , the following engel - type theorem is an immediate consequence of the same theorem for usual lie algebras ( * ? ? ?
* th 6.4.1 ) .
suppose @xmath3 is a commutative @xmath141-algebra , @xmath13 is a tensor category over @xmath3 , and @xmath112 is a lie @xmath13-algebra satisfying the linearised engel identity . then @xmath112 satisfies the linearised nilpotency . nevertheless , in our view , it is interesting to understand examples of @xmath13-algebras , in particular , what identities they satisfy .
for instance , if @xmath2 is a rigid object in a tensor category @xmath13 then @xmath144 is an associative @xmath13-algebra , so - called _ the internal endomorphism algebra_. the identities of @xmath145 in the free associative algebra @xmath146 form a @xmath142-ideal @xmath147 .
it is interesting to study it @xcite . given a rigid object @xmath2 in @xmath13 , what is the minimal degree of an element of @xmath148 ?
if @xmath13 is the category of vector spaces , @xmath2 is an @xmath86-dimensional vector space then @xmath145 is the algebra of @xmath149 matrices , whose minimal identity has degree @xmath150 by the amitsur - levitzki theorem @xcite .
if @xmath13 is the category of supervector spaces ( this is the same as superbundles on a point , see section [ superbundle ] ) , @xmath2 is an @xmath151-dimensional supervector space then @xmath145 is the matrix superalgebra whose minimal identity is conjectured to be of the degree @xmath152 @xcite .
one reason the matrix superalgebras are of interest is that in characteristic zero they generate all prime varieties of associative algebras @xcite .
is it conceivable that algebras @xmath145 generate all varieties ? give an example of a commutative @xmath141-algebra @xmath3 and a @xmath142-ideal in @xmath153 that is not an ideal of identities of @xmath145 for any rigid object @xmath2 in any tensor category @xmath13 over @xmath3 .
we finish this section explaining the categorical analogue of the grassmann envelope used to study the varieties of superalgebras . _
the global sections functor _
@xmath154 is defined by @xmath155 .
if @xmath28 is an algebra in @xmath13 then @xmath156 is a @xmath17-algebra in @xmath17-modules .
the following proposition is evident .
[ decat_alg ] suppose @xmath13 is a tensor category , @xmath28 is a @xmath13-algebra , @xmath157 is an associative commutative @xmath13-algebra .
then @xmath158 is a @xmath159-algebra of the same signature as @xmath28 .
moreover , @xmath158 satisfies all identities of @xmath28 .
the grassmann algebra @xmath160 in countably many variables is an associative commutative algebra in the category @xmath13 of supervector spaces .
if @xmath28 is a superalgebra then @xmath161 is the grassmann envelope of @xmath28 .
the identities of @xmath161 are precisely the identities of @xmath28 in @xmath13 .
can this be carried out in any category ?
let @xmath13 be a cocomplete tensor category with split idempotents is the projection operator @xmath162 . ] .
does there exist a commutative associative algebra @xmath163 in @xmath13 such that for any @xmath13-algebra @xmath28 all identities of @xmath164 hold in @xmath28 ?
following vogel @xcite , a lie algebra @xmath165 is _ quasisimple _ if it can be extended to a metric lie algebra @xmath166 and the natural map @xmath167 is an isomorphism . we discuss how it is related to the usual simplicity in the next section . if the @xmath3-tensor category @xmath13 is abelian , we can talk about subobjects , quotients , kernels , ideals , etc . in the usual way .
in particular , there is the usual notion of a simple lie algebra .
however , the category being abelian is not required to introduce simplicity , although the utility of this notion is unclear .
neither is it clear whether this definition is useful for other classes of algebras .
recall that a morphism @xmath168 is _ a monomorphism _ if it can be cancelled on the left , i.e. , @xmath169 implies @xmath170 for all @xmath171 .
a lie algebra @xmath112 in @xmath13 is _ simple _ if @xmath172 ( notice that if a zero object exists , it is unique up to a canonical isomorphism ) and every algebra homomorphism @xmath173 is a monomorphism or the zero . notice the fine difference with the usual simple lie algebras in vector spaces : the latter definition traditionally excludes the one - dimensional lie algebra while our definition does not . in general , there is no relation between simple and quasisimple lie algebras .
let @xmath3 be a commutative ring , not a field , @xmath174 .
it admits a proper ideal @xmath68 .
then the lie algebra @xmath175 ( in the category of @xmath3-modules ) is quasisimple : the trace form @xmath176 gives a metric structure with a coform @xmath177 it is not simple because the quotient homomorphism @xmath178 is neither zero , nor a monomorphism .
here is another example of a quasisimple non - simple lie algebra .
let @xmath2 be a set of simple finite dimensional complex lie algebras , @xmath179 the group coalgebra .
a @xmath157-comodule is just an @xmath2-graded vector space @xmath180 .
we consider the tensor category @xmath13 of @xmath157-comodules with the standard tensor coproduct as a tensor product : @xmath181 . now
@xmath182 is a metric lie @xmath13-algebra under the direct sums of the products , the forms and the coforms @xmath183 observe that @xmath112 is quasisimple but , if @xmath2 contains at least two elements , not simple since every @xmath184 is a proper ideal of @xmath112 resulting in the quotient homomorphism . in the opposite direction
, @xmath185 is a simple lie algebra in the category of real vector spaces but it is not quasisimple because @xmath186 .
nevertheless , sometimes there is a relation .
[ simple_in_vs ] let @xmath13 be the category of vector spaces over an algebraically closed field @xmath3 of characteristic zero .
then quasisimple lie @xmath13-algebras are exactly the finite dimensional simple lie algebras . a finite dimensional simple lie algebra @xmath112 is quasisimple because the killing form is non - degenerate and @xmath187 is an isomorphism by the schur lemma and algebraic closeness of @xmath3 . in the opposite direction ,
the space of symmetric invariant form @xmath188 is one - dimensional since it is a subspace of @xmath189 . by the theorem of bajo and benayadi @xcite
, @xmath112 is simple .
it would be interesting to find some other categories where an analogue of theorem [ simple_in_vs ] holds or where the quasisimple lie algebras admit a meaningful classification .
let us pose two precise questions .
let @xmath163 be an affine group scheme over a field @xmath3 of characteristic zero .
let @xmath13 be the category of rational @xmath163-modules .
what are simple and quasisimple lie algebras in @xmath13 ? to get a feel of this question ,
let us consider an absolutely irreducible finite dimensional @xmath163-module @xmath190 as a lie algebra with the zero multiplication in @xmath13 . according to our definition
, @xmath190 is simple . on the other hand
, @xmath190 is quasisimple if and only if @xmath191 .
let @xmath13 be the category of supervector spaces - modules as a monoidal category , but not as a tensor category .
] over an algebraically closed field @xmath3 of characteristic zero . is a quasisimple lie @xmath13-algebra a finite dimensional simple lie superalgebra ?
finite dimensional simple lie superalgebras are classified by kac @xcite .
simple lie algebras and simple lie superalgebras of types @xmath192 , @xmath193 , @xmath194 , @xmath195 , @xmath196 , @xmath197 , @xmath198 , and @xmath199 have nondegenerate killing forms ( * ? ? ?
* th . 1 ) .
hence , they are quasisimple lie @xmath13-algebras .
the remaining simple lie superalgebras of types @xmath200 , @xmath201 , @xmath202 , @xmath203 , and @xmath204 as well as cartan type superalgebras have zero killing forms ( * ? ? ?
2.4.1 ) , although there may be an invariant form distinct from the killing form .
for instance , this happens in types @xmath203 and @xmath204 .
her is a partial result towards this question .
[ qs_super ] let @xmath112 be a quasisimple lie algebra in the category of supervector spaces over a field @xmath3 . then @xmath112 is perfect , the centre of @xmath112 is zero , every minimal ideal of @xmath112 is abelian .
let @xmath68 be a minimal ideal .
its orthogonal complement @xmath205 is also an ideal .
hence , the intersection @xmath206 is zero or @xmath68 . if @xmath207 , then @xmath208 and we get nontrivial idempotents in @xmath209 .
hence @xmath210 , so @xmath211 .
this means that @xmath212,a)=(x,[y , a])=0 $ ] for all @xmath213 , @xmath214 .
as the form is nondegenerate , @xmath215=0 $ ] and @xmath68 is abelian .
now @xmath216^\perp = \ { a\in{{\mathfrak g}}\mid \forall x , y\in { { \mathfrak g}}\ ( [ x , y],a)=(x,[y , a])=0\ } = \ { a\in{{\mathfrak g}}\mid \forall y\in { { \mathfrak g}}\ [ y , a]=0\ } = z({{\mathfrak g}})$ ] .
we conclude that @xmath112 is perfect and @xmath217 .
indeed , otherwise there exists a nonzero bilinear form @xmath218 on @xmath219 $ ]
. extend it to a form @xmath220 on @xmath112 : @xmath221,y+[{{\mathfrak g}},{{\mathfrak g}}])$ ] .
the form @xmath220 is invariant ( @xmath222,z)=0=\widetilde{\alpha } ( x,[y , z])$ ] ) , nonzero and degenerate .
this contradicts the one - dimensionality of @xmath223 . using proposition [ qs_super ] and cartan s criteria ( * ? ? ?
* and 5.1 ) , we can furnish an alternative proof of theorem [ simple_in_vs ] . consider @xmath112 as a superalgebra with zero odd part , hence proposition [ qs_super ] applies . since @xmath112 is not solvable , the killing form is nonzero .
so the killing form is a multiple of the non - degenerate form , and itself nondegenerate .
thus , @xmath112 is semisimple .
it is simple since @xmath224 .
this proof fails for the superalgebras because of the lack of the cartan criteria .
the universal metric lie algebra appears in the study of vassiliev invariants @xcite .
the relevant symmetric monoidal category @xmath13 is a prop .
the hom - set @xmath225 is the @xmath141-vector space of jacobi diagrams @xcite .
recall that an @xmath226-jacobi diagram is a compact curve @xmath2 such that the boundary of @xmath2 is the set @xmath227 , @xmath2 has finitely many trivalent singular points , i.e. points @xmath228 with a neighbourhood @xmath229 such that @xmath230 is a union of three lines , for each trivalent singular point @xmath228 a cyclic ordering on the components of @xmath230 is fixed .
here is an example of the jacobi diagram in @xmath231 : ( 50,50)(-140,0 ) ( 5,40)(0,0)@xmath232 ( 5,23)(0,0)@xmath233 ( 15,23)(0,0 ) .
( 5,6)(0,0)@xmath234 ( -63,34)(0,0)@xmath235 ( -63,11)(0,0)@xmath236 ( -18,12)(0,0)@xmath237 the external dashed borders have no significance .
this diagram has five trivalent singular points , four of them have the standard counterclockwise ordering , while the remaining vertex marked with @xmath237 has the opposite clockwise ordering .
the point of transversal intersection has no significance : it is actually two distinct points of the curve which coincided after an immersion into a plane ( a.k.a drawing ) . now @xmath225 is the quotient space of the @xmath141-span of all @xmath226-jacobi diagram subject to the @xmath238-relation and the @xmath239-relation : ( 250,50)(-30,0 ) ( -246,25)(0,0)@xmath240 ( -104,25)(0,0)@xmath241 ( -50,25)(0,0)@xmath240 ( -211,21)(0,0)@xmath237 ( -182,25)(0,0)@xmath242 , ( 10,25)(0,0)@xmath242 . the dashed circles bound neighbourhoods in jacobi diagrams @xmath243 , @xmath244 for the first relation and @xmath245 , @xmath246 , @xmath247 for the second one , which are identical except for these neighbourhoods .
the relations mean that @xmath248 the composition and the tensor product of morphisms is done by `` stacking '' the boxes : ( 250,60)(-30,0 ) ( -237,31)(0,0)@xmath249 ( -161,31)(0,0)@xmath249 ( -64,31)(0,0)@xmath249 ( -15,13)(0,0)@xmath249 ( -287,31)(0,0)@xmath2 ( -191,31)(0,0)@xmath2 ( -110,31)(0,0)@xmath2 ( -15,41)(0,0)@xmath2 ( -262,30)(0,0)@xmath250 ( -88,30)(0,0)@xmath251 ( -140,30)(0,0 ) ; ( -213,30)(0,0)@xmath252 ( -40,30)(0,0)@xmath252 thus , the tensor product of morphisms is just a union while the composition is gluing along the corresponding part of the boundary .
the symmetric group @xmath31 embeds in a way that @xmath253 is a union of @xmath86 intervals connecting @xmath77 with @xmath254 : ( 200,70)(-150,0 ) ( 5,63)(0,0)@xmath232 ( 5,48)(0,0)@xmath233 ( 5,20)(0,0)@xmath255 ( 5,3)(0,0)@xmath86 ( 17,32)(0,0)@xmath256 ( -95,63)(0,0)@xmath235 ( -103,47)(0,0)@xmath257 ( -103,33)(0,0)@xmath258 ( -103,16)(0,0)@xmath259 ( -70,20)(0,0)@xmath255 ( -15,27)(0,0)@xmath255 ( -140,34)(0,0)@xmath260 ( 44,34)(0,0 ) .
the @xmath13-algebra @xmath261 where ( 250,50)(-40,0 ) ( -290,25)(0,0)@xmath235 ( -63,34)(0,0)@xmath235 ( -63,12)(0,0)@xmath236 ( -110,35)(0,0)@xmath232 ( -110,12)(0,0)@xmath233 ( -222,34)(0,0)@xmath232 ( -222,12)(0,0)@xmath233 ( -315,23)(0,0)@xmath262 ( -191,23)(0,0 ) , @xmath263 ( -88,23)(0,0 ) , @xmath264 is a metric lie algebra ( * ? ? ?
the reader can easily verify this .
jacobi identity and anticommutativity follow from the as - relation and the ihx - relation .
the axioms of a metric object , as well as invariance of the form and co - form are evident on a drawing .
the universal property of this metric lie algebra is more subtle ( * ? ? ?
we state it as a separate proposition .
given a commutative @xmath141-algebra @xmath3 and a metric lie @xmath3-algebra @xmath265 , there exists a unique specialization @xmath266 such that @xmath267 is the @xmath141-algebra structure on @xmath3 , @xmath268 is a tensor functor , defined on objects by @xmath269 , and uniquely determined on morphisms by the requirements that @xmath270 , @xmath271 , @xmath272 , @xmath273 is the composition of associativity constraints in @xmath46 , @xmath274 , the identity morphism of @xmath275 , @xmath276 , the identity morphism of @xmath112 .
the specialization @xmath277 is an example of isospecialization .
its constituent is a functor @xmath278 that turns @xmath112 into an algebra over the prop @xmath279 . in the language of props
@xmath279 is the universal metric lie algebra over @xmath3 : @xmath279-algebras are the same as metric lie @xmath3-algebras ( * ? ? ?
in the rest of the section we will show that @xmath280 is not quasisimple .
consider @xmath281 , the @xmath141-linear span of all those diagrams in @xmath282 that contain precisely @xmath77 connected components .
the @xmath238-relation and @xmath239-relation do not change the number of connected components .
hence the spaces @xmath281 for different @xmath77 have zero intersections , while @xmath282 splits into a direct sum @xmath283 .
notice that @xmath284 if and only if @xmath285 : it is spanned by the empty diagram .
the tensor products preserve this grading , but the compositions , in general , destroy it .
let us summarize properties of the grading .
[ lambda ] the following statements hold .
@xmath286 for all @xmath77 , @xmath287 , @xmath86 , @xmath288 .
@xmath289 for all @xmath77 , @xmath287 , @xmath86 , @xmath288 . under this grading
the scalar ring @xmath290 is a graded algebra , isomorphic to the symmetric algebra of @xmath291 . under this grading
each @xmath282 is a graded @xmath17-module .
the tensor product of morphisms is a disjoint union .
hence , the number of components of the tensor product is the sum , and ( 1 ) follows .
a similar argument proves ( 2 ) .
( 4 ) immediately follows from ( 1 ) .
it follows from ( 1 ) that @xmath292 is a graded commutative algebra .
thus , there is a natural homomorphism @xmath293 of graded algebras .
it is surjective because each diagram is a product of its components .
let us show that @xmath294 is injective .
consider @xmath295 , all connected jacobi diagrams without boundary before the @xmath238-relation and the @xmath239-relation are imposed .
there are natural homomorphisms of graded algebras @xmath296 , @xmath297 making the diagram ( 150,60)(0,0 ) ( 0,60)@xmath298 ( 105,60)@xmath292 ( 70,63)(1,0)30 ( 80,55)@xmath294 ( 13,8)@xmath299 $ ] ( 42,19)(0,1)35 ( 27,35)@xmath296 ( 58,19)(3,2)56 ( 91,32)@xmath297 commutative . the kernel of @xmath296 is the ideal @xmath300 generated by all @xmath301 from the @xmath238-relation and all @xmath302 from the @xmath239-relation , where @xmath303 , @xmath304 and @xmath305 are connected . similarly , the kernel of @xmath297 is the ideal @xmath306 generated by all @xmath307 from the @xmath238-relation and all @xmath308 from the @xmath239-relation , where @xmath309 , @xmath310 and @xmath311 are are no longer necessarily connected .
both relations operate on just one component , that is , if @xmath307 is an @xmath238-relation , @xmath312 and @xmath313 where @xmath314 is the component affected by the @xmath238-relation and @xmath315 is the union of the remaining components .
thus , @xmath316 .
similarly , @xmath317 for all @xmath239-relations .
it follows that @xmath318 and @xmath294 is injective .
since connected jacobi diagrams span @xmath319 , one can choose a basis @xmath320 among them .
it follows that @xmath290 is isomorphic the polynomial algebra @xmath321 $ ] on the chosen connected @xmath322-jacobi diagrams such as ( 250,50)(-60,0 ) ( -160,23)(0,0)@xmath323 ( -71,23)(0,0 ) , @xmath324 ( 16,23)(0,0 ) , @xmath325 ( 60,26)(0,0)@xmath326 , while @xmath327 is a graded @xmath17-module .
the homomorphism @xmath328 is the action on the identity element @xmath329 . since @xmath330 and @xmath331 , any connected non - identity diagram from @xmath332 is not in the image of the map @xmath333 . in particular , the following element @xmath334 is not in the ( -71,23)(0,0)@xmath335 ( 64,23)(0,0 ) .
( 34,26)(0,0)@xmath336 the symmetric group @xmath31 acts on the vector space @xmath338 by permuting the inputs . as a vector space
, vogel s ring @xmath337 is @xmath339 , the skew invariants with respect to the sign character @xmath340 of @xmath341 .
the multiplication is via insertion of one diagram into any trivalent point of the second diagram : ( 250,50)(-40,0 ) ( -40,18)(0,0)@xmath249 ( -110,23)(0,0)@xmath249 ( -10,40)(0,0)@xmath2 ( -166,40)(0,0)@xmath2 ( -148,23)(0,0)@xmath342 ( -69,23)(0,0)@xmath252 ( 13,23)(0,0 ) . observe that connectedness and skew - invariance of elements of @xmath343 are crucial for this product to be well - defined .
once a trivalent singular point is chosen , there are six ways to insert a diagram from @xmath343 .
skew - invariance removes this dependency .
if two singular points are connected by an edge , one can see that insertions into each of these two points give the same result .
consequently , the connectedness eliminates the dependency of insertion on the choice of a singular point .
vogel proves these facts ( * ? ? ?
* prop 3.2 ) . at the same time he establishes that the insertion defines a structure of a @xmath337-module on the linear span of connected diagrams with a non - empty set of trivalent singular points @xmath344 . while associativity of @xmath337 is obvious , commutativity requires a subtle argument ( * ? ? ? * prop 4.8 ) .
notice that vogel works over any commutative coefficient ring @xmath345 and proves the identity @xmath346 in @xmath347 . in our case @xmath348 , therefore the ring @xmath337 is commutative . here
are some elements of @xmath337 : ( 250,50)(-40,0 ) ( -43,27)(0,0)@xmath255 ( -37,24)(0,0)@xmath86 ( -315,23)(0,0)@xmath349 ( -191,23)(0,0 ) , @xmath350 ( -88,23)(0,0 ) , @xmath351 ( 13,23)(0,0 ) .
let us note that @xmath352 and @xmath353 @xcite .
the ring @xmath337 is naturally graded : @xmath354 is spanned by the diagrams with @xmath355 trivalent singular points .
the poincare series of @xmath337 is not known .
the ring @xmath337 is closely related to a subalgebra @xmath356 \oplus \omega { { \mathbb q}}[z_1,z_2,z_3]^{s_3}$ ] of @xmath357^{s_3}$ ] where @xmath358 and @xmath359 , @xmath360 , @xmath361 are the elementary symmetric polynomials .
kneissler constructs further elements @xmath362 by @xmath363 and a homomorphism of graded algebras @xcite @xmath364 vogel constructs a polynomial @xmath365 such that @xmath366 and conjectures that @xmath367 @xcite .
there is a canonical invariant non - degenerate form ( the killing form ) @xmath368 on a simple complex lie algebra @xmath112 .
it determines a canonical isospecialization @xmath369 that , in its turn , defines a canonical ring homomorphism @xmath370 .
it is sensible to call it _ the vogel character of lie algebra _ @xmath112 .
let us clarify its origin @xcite .
an element @xmath371 gives a @xmath112-invariant map @xmath372 .
the space of such maps is one - dimensional , hence , @xmath373 for some complex number @xmath218 . in the notation of the last section
the map @xmath374 is a ring homomorphism .
it is obvious that @xmath375 is linear .
let @xmath376 be the result of removal of a neighbourhood of a singular point from @xmath249 .
now one can represent products in vogel ring via compositions : @xmath377 which implies that @xmath378 and , finally , @xmath379 .
vogel has realized how to combine the character @xmath380 and the isospecialization @xmath381 into a single structure by forcing the action of @xmath337 on the category @xmath13 @xcite .
the new tensor category @xmath45 is a quotient of @xmath382 by the vogel relations : ( 250,50)(-40,0 ) ( -302,38)(0,0)@xmath2 ( -208,38)(0,0)@xmath2 ( -240,17)(0,0)@xmath383 ( -118,38)(0,0)@xmath249 ( -10,38)(0,0)@xmath249 ( -272,23)(0,0)@xmath384 ( -185,23)(0,0)@xmath385 ; ( -86,23)(0,0)@xmath386 ( 10,23)(0,0)@xmath387 . thus , @xmath45 is also a prop with morphisms @xmath388 where @xmath389 is the @xmath337-span of all @xmath390 where @xmath391 , @xmath243 and @xmath244 differ as on the diagrams in the first relation , @xmath245 and @xmath246 differ as on the diagrams in the second relation .
[ vog_m ] @xmath45 is a tensor category over @xmath392 $ ] .
@xmath393 is a quasisimple lie @xmath45-algebra .
observe that the tensor product and the composition of cosets @xmath394 are well defined . to see this
, it suffices to establish that the collection of subspaces @xmath395 is a tensor ideal , that is , @xmath396 by @xmath337-linearity , we can assume that @xmath397 , @xmath398 and @xmath399 .
the argument used in proposition [ lambda ] is based on the local nature of the relations .
it works here as well .
let @xmath400 where @xmath243 and @xmath244 are diagrams related by the first vogel relation .
then @xmath401 and @xmath402 are also related by the first vogel relation , and @xmath403 .
the same argument works for the second vogel relation and the tensor products .
thus , @xmath45 is a tensor category . to compute the scalars ,
let us recall that the span of the connected diagrams with at least one singular point @xmath404 is a @xmath337-module ( * ? ? ?
* prop 3.2 ) .
proposition [ lambda ] implies that @xmath405 \otimes_{{\mathbb q}}s ( \hom^s_{{{\mathcal c } } } ( 0,0)),\ ] ] because @xmath406 is the only connected diagram from @xmath407 without singular points . the element @xmath408 is a free generator of the @xmath337-module @xmath409 ( * ? ? ?
* cor 4.6 ) , hence , @xmath410
\otimes_{{\mathbb q}}s ( \lambda \cdot \theta).\ ] ] let @xmath411 be the @xmath337-span of the first vogel relations inside @xmath412 .
since the first vogel relation asserts @xmath337-linearity on @xmath413 , @xmath414 \otimes_{{\mathbb q}}\mbox{sym}_\lambda ( \lambda \cdot \theta)/ ( \theta -2t\delta ) \cong \lambda [ \delta].\ ] ] evidently , the second vogel relation has no further effect and @xmath415,\ ] ] proving the first statement .
since @xmath280 is a metric lie algebra , @xmath0 is a metric lie algebra .
furthermore , @xmath416 because @xmath334 is a free generator of the @xmath337-module @xmath417 ( * ? ? ?
* cor 4.6 ) and @xmath418 , @xmath419 ( * ? ? ?
finally , @xmath420 we say that a lie @xmath13-algebra @xmath112 is _
@xmath190-simple _ if it is quasisimple and a specialization @xmath421 extends to a specialization @xmath422 . clearly , all this definition requires is a ring homomorphism @xmath423 \rightarrow { { \mathbb k}}_{{\mathcal c}}$ ] satisfying the vogel relations .
existence of such a homomorphism is not clear to us , in general .
vogel gives a categorical criterion for a quasisimple lie algebra to be @xmath190-simple @xcite but it is of limited use . to prove that a quasisimple lie superalgebra @xmath112 is @xmath190-simple @xcite it is easier to use the homomorphism @xmath375 constructed at the end of section [ vrl ] ( notice that @xmath424 in any category ) . in the final chapter we use the same strategy of constructing a ring homomorphism explicitly while the vogel categorical criterion fails .
so far , all known to us quasisimple lie algebras are @xmath190-simple , hence , the following question is interesting .
let @xmath13 be a tensor category .
is a quasisimple lie @xmath13-algebra @xmath190-simple ? finally , let us discuss the connection between the vogel plane and the vogel ring . by _ the vogel plane _
one commonly understands the weighted projective plane @xmath425 @xcite . a point .
] @xmath426 , where @xmath218 , @xmath427 , @xmath428 are nonzero eigenvalues of casimir operator on the symmetric square of the adjoint representation @xmath429 , corresponds to a simple complex lie algebra @xmath112 .
when the invariant form changes , the casimir operator changes by a scalar , hence , the triple @xmath430 is also defined only up to a scalar .
this construction has an issue with exceptional ( according to deligne ) lie algebras by deligne @xcite . a lie algebra of type @xmath431 , @xmath432 , @xmath433 , @xmath434 or @xmath435 admits only two such eigenvalues : @xmath218 and @xmath427 .
nevertheless , it is sensible to define @xmath436 where @xmath406 is the casimir eigenvalue on the adjoint representation , because the identity @xmath437 holds in exceptional lie algebras .
there is a more series issue with @xmath438 , it admits only one eigenvalue @xmath218 . in this case , one assigns a whole line @xmath439 .
now let us consider a @xmath190-simple lie algebra @xmath112 over a ring @xmath3 with vogel character @xmath423 \rightarrow { { \mathbb k}}$ ] .
homomorphisms of graded rings @xmath440\hookleftarrow { { \mathbb q}}[z_1,z_2,z_3]^{s_3}\hookleftarrow \widetilde{\lambda } \xrightarrow{\varphi } \lambda\ ] ] give another homomorphism @xmath441 .
rings @xmath357^{s_3}$ ] and @xmath442 have a common ring of fractions @xmath357^{s_3}[\omega^{-1 } ] = \widetilde{\lambda } [ \omega^{-1}]$ ] , hence , if @xmath443 is invertible in @xmath3 , then one gets a graded homomorphism @xmath444^{s_3 } \rightarrow { { \mathbb k } } , \ \
\widehat{\theta}_{{\mathfrak g}}(f ) = \frac{\widetilde{\theta}_{{\mathfrak g}}(f\omega)}{\widetilde{\theta}_{{\mathfrak g}}(\omega)}.\ ] ] examples of algebras with undefined @xmath445 are @xmath446 and k3-surfaces in the next chapter . in both cases @xmath447 .
another consequence of invertibility of @xmath443 is that the @xmath448-orbit @xmath449 $ ] turns out to be a point of the weighted projective plane @xmath450 over the ring @xmath3 .
recall that the multiplicative group @xmath448 acts on the set @xmath451^{s_3 } , { { \mathbb k}})$ ] via the grading of the ring @xmath357^{s_3}$ ] : @xmath452_n^{s_3 } , \
\theta \in \hom ( { { \mathbb q}}[z_1,z_2,z_3]^{s_3 } , { { \mathbb k}}),\ ] ] and the set of @xmath3-points @xmath453 is a subset of the set of orbits .
note that @xmath445 depends on the metric lie algebra @xmath112 .
a different choice of a metric associated is linked by a scalar @xmath454 , resulting in the homomorphism @xmath455 .
thus , the orbit @xmath449\in{{\mathbb p}}_{1,2,3}^2 ( { { \mathbb k}})$ ] depends only on the lie algebra itself , but not on its metric structure .
a holomorphic manifold @xmath2 admits an associated tensor category , crucial for kapranov s approach to rozansky - witten invariants @xcite .
we attach a slightly smaller category than kapranov .
we explain the difference after we explain the construction .
the objects in the category @xmath456 of vector superbundles are locally free coherent sheaves @xmath457 .
the tensor product is usual : @xmath458 a locally free coherent sheaf @xmath459 gives to two objects : even @xmath460 and odd @xmath461 .
the unit object is the even trivial line bundle @xmath462 .
the symmetric braiding on @xmath456 is the usual superbraiding @xmath463 where @xmath464 , @xmath465 are homogeneous of degree @xmath133 or @xmath466 , i.e. local sections of @xmath467 and @xmath468 , correspondingly .
the category @xmath456 is a full subcategory of the 2-periodic derived category @xmath473 of coherent sheaves on @xmath2 @xcite .
a superbundle @xmath474 corresponds to a complex @xmath475 \oplus{{\mathcal f}}_1[2n+1])$ ] with zero differentials . in the literature
some `` larger '' categories are sometimes considered .
first , one can get more morphisms in by considering all extensions : @xmath476 resulting in a category @xmath477 : @xmath478 is a subcategory of @xmath477 with the same objects but larger sets of morphisms .
the category @xmath477 is a full subcategory of the unfolded derived category @xmath479 , whose objects and tensor products are from @xmath480 , while morphisms account for all extensions : @xmath481).\ ] ] for us all these larger categories are irrelevant since all the necessary objects and morphisms for the kapranov theorem are in @xmath456 .
let @xmath482 be the tangent sheaf on @xmath2 , @xmath483 the sheaf of differential operators with holomorphic coefficients of order less than @xmath86 .
we have an exact sequence of @xmath484-bimodules @xmath485 notice that on both @xmath486 and @xmath482 the right and the left actions of @xmath486 coincide but on @xmath487 they are different . given a locally free coherent sheaf @xmath459 , we get a new exact sequence by tensoring with it @xmath488 this is extension is the atiyah class @xmath489 let us remark that the standard atiyah class is actually @xmath490 @xcite .
the following theorem has essentially been proved by kapranov @xcite , but a reader may benefit by looking at later reviews @xcite .
kapranov s original proof establishes this theorem in the bigger category @xmath495 @xcite . we just need to point out that everything in kapranov s proof happens in either odd or even extensions , so the theorem actually holds in @xmath456 .
in particular , observe that @xmath494 is skew - symmetric and @xmath496 is odd , hence @xmath494 is symmetric in @xmath456 .
let @xmath497 ( or @xmath498 , or @xmath499 ) be the @xmath86-th symmetric ( or tensor , or exterior ) power of a locally free coherent sheaf @xmath459 .
let us make a couple of general observations before addressing quasisimplicity of @xmath1 .
first , observe that @xmath500 for all @xmath501 ( * ? ? ?
* cor 4.39 ) . secondly , observe for the representations of @xmath502 . ]
that for a holomorphic symplectic manifold @xmath2 @xmath503 with the trace map splitting the morphism @xmath504 , in particular , on an open set @xmath505 @xmath506 the scalars of the category @xmath456 are @xmath508 the naturality of the atiyah class means that every homomorphism in @xmath456 is a homomorphism of representations of @xmath1 .
hence , @xmath509 and the natural map @xmath510 is the identity on the third component .
the even cohomology is easy to control in the case of k3-surfaces .
this is in sharp contrast to the general behaviour of the symmetric plurigenus @xmath511 that is not a topological invariant of a complex manifold @xmath2 @xcite .
the crucial argument for the next theorem has been explained to us by a. kuznetsov . by theorem [ manifold_simple ]
it suffices to prove that @xmath512 vanishes because @xmath513 for a surface .
suppose @xmath514 .
then we have two sections of @xmath515 the skew symmetric identity map @xmath516 and some symmetric section @xmath517 .
the determinant @xmath518 is a global function on @xmath2 for each @xmath519 .
thus , it is constant and one can choose @xmath520 such that @xmath521 .
we conclude that the rank of @xmath522 is @xmath232 at a generic point .
furthermore , since @xmath68 is skew - symmetric and @xmath523 is symmetric , @xmath524 does not vanish , so the rank is @xmath232 at each point .
thus , the image of @xmath524 is an invertible sheaf @xmath525 .
let us show that @xmath525 must be trivial .
if @xmath2 is not algebraic then the only line bundle on @xmath2 is trivial .
if @xmath2 is algebraic then @xmath482 is semistable @xcite ( cf .
* ch . 7.4 ) ) and @xmath526 for any ample divisor @xmath527 .
the dual of the natural map @xmath528 is an embedding @xmath529 , hence @xmath530 .
thus , @xmath531 for any ample divisor @xmath527 , proving that @xmath525 is trivial . since @xmath525 is trivial , @xmath525 and
consequently @xmath482 has a nonzero section .
this contradicts @xmath532 .
the specialization @xmath533 is known as the rozansky - witten invariants ( a.k.a weight system ) of a holomorphic symplectic manifold @xmath2 .
it is computed up to various degrees of explicitness for many concrete @xmath2 @xcite .
let @xmath534 be the holomorphic symplectic form on @xmath2 , @xmath535 its conjugate form , and @xmath536 \in h^2(x,{{\mathcal o}}_x)$ ] the class corresponding to @xmath494 under the natural isomorphism @xmath537 .
using @xmath494 , one gets natural embeddings @xmath538 so that for @xmath539 with @xmath77 singular trivalent points is always even . ]
@xmath540 in particular ,
if @xmath2 is a k3-surface , then @xmath541 whenever @xmath428 has at least three trivalent singular points . for the diagrams with @xmath542 we have @xcite @xmath543 , \
{ { { \mathrm r}{\mathrm w}}}_2 ( \theta ) = 48[\overline{\varsigma } ] \in h^2(x,{{\mathcal o}}_x ) . \ \ ] ] by theorem [ surface_simple ] the lie algebra @xmath1 is quasisimple and @xmath544 / ( z^2 ) \
\mbox { where } \
z= [ \overline{\varsigma } ] \in h^2(x,{{\mathcal o}}_x).\ ] ] a graded homomorphism @xmath545 / ( z^2 ) , \ \
\theta_x(t ) = -24 z\ ] ] is uniquely determined since @xmath546 and @xmath547 .
we claim that @xmath548 defines a required specialization .
indeed , most of the relations defining @xmath45 hold in @xmath456 for the trivial reason : both sides are zero .
the relations that @xmath549 implies @xmath550 . ] @xmath550 and @xmath549 hold as both sides become @xmath551 and @xmath552 .
theorem [ surface_vsimple ] proves westbury s conjecture for a k3-surface .
westbury has conjectured it for any compact irreducible holomorphic symplectic manifold @xmath249 .
we finish by stating this conjecture more carefully as a series of questions .
the quasisimplicity of @xmath553 can be established by theorem [ manifold_simple ] and boils down to the following question .
, a manifold @xmath249 with a positive answer to question 7 must have @xmath554 .
if @xmath555 , then one can use its three kahler structures to construct two linearly independent classes in @xmath556 , in particular , @xmath557 . ]
if @xmath559 then the scalars of @xmath560 are @xmath561 / ( z^n ) $ ] where @xmath562 \in h^2(y,{{\mathcal o}}_y ) .
$ ] thus , the @xmath190-simplicity of @xmath553 boils down to some explicit identities on the rozansky - witten invariants .
does there exists a @xmath190-simple lie algebra @xmath565 over @xmath566 $ ] that specializes to @xmath567}}$ ] for all @xmath2 and @xmath86 where @xmath568}$ ] is the hilbert scheme of @xmath86 points on a k3-surface @xmath2 , so that the specialization @xmath569}}$ ] factorises as @xmath570}}$ ] ? | we study algebras defined by identities in symmetric monoidal categories .
our focus is on lie algebras . besides usual lie algebras , there are examples appearing in the study of knot invariants and rozansky - witten invariants . our main result is a proof of westbury s conjecture for k3-surface : there exists a lie algebra homomorphism from vogel s universal simple lie algebra to the lie algebra describing the rozansky - witten invariants of a k3-surface .
most of the paper involves setting up a proper language to discuss the problem and we formulate nine open questions as we proceed . in 1996
deligne made a conjecture that exceptional lie algebras formed a series @xcite .
he proposed a possible explanation : he expected an object specializing to all exceptional lie algebras .
influenced by this conjecture , vogel proposed such an object in 1999 @xcite .
besides exceptional lie algebras , it specializes to all simple ( complex finite dimensional ) lie algebras as well as some lie superalgebras .
the object was named vogel s universal simple lie algebra @xmath0 . with vogel s paper still unpublished the object remains a mystery .
it is a subject of several recent studies @xcite .
the aim of this paper is to describe another specialization of @xmath0 . in a private conversation
westbury has asked the author whether @xmath0 specializes to @xmath1 , a certain lie algebra associated by kapranov @xcite to an irreducible holomorphic symplectic manifold @xmath2 in the study of rozansky - witten invariants @xcite . by _
westbury s conjecture _ we understand existence of such a specialization .
we give an affirmative answer to westbury s conjecture in the case of a k3-surface .
it remains open in the general case .
most of the paper is devoted to setting up a machinery just to ask the question rigorously .
there are several levels of generality in which the setup can be laid out .
as opposed to a more general operadic language of hinich and vaintrob @xcite , we choose a more elementary language in the spirit of sawon @xcite to benefit a wider audience .
here is a detailed description of the paper .
section 1 contains all necessary definitions and facts about the categories . in section 2
we discuss algebras in the categories .
we speculate on possible utility of such algebras in the study of identities . in section 3
we study the lie algebras .
we formulate three different notions of simplicity .
we introduce vogel s universal lie algebra . in section 4
we investigate the lie algebras that appear in the study of the rozansky - witten invariants and give an affirmative answer to westbury s conjecture in the case of a k3-surface .
the author expresses especial gratitude to b. westbury for formulating the conjecture and his interest in the work and to a. kuznetsov for explaining the stability of the tangent bundle on a k3-surface .
the author would like to thank a. baranov , v. kac , d. panyushev , a. rosly and j. sawon for valuable discussions and useful information . |
we consider the helmholtz equation of the form @xmath0 where @xmath1 is the wave number , @xmath2 represents a harmonic source , @xmath3 is a given data function , and @xmath4 is a spatial function describing the dielectric properties of the medium . here @xmath5 is a polygonal or polyhedral domain in @xmath6 . under the assumption that the time - harmonic behavior is assumed
, the helmholtz equation ( [ pde ] ) governs many macroscopic wave phenomena in the frequency domain including wave propagation , guiding , radiation and scattering .
the numerical solution to the helmholtz equation plays a vital role in a wide range of applications in electromagnetics , optics , and acoustics , such as antenna analysis and synthesis , radar cross section calculation , simulation of ground or surface penetrating radar , design of optoelectronic devices , acoustic noise control , and seismic wave propagation .
however , it remains a challenge to design robust and efficient numerical algorithms for the helmholtz equation , especially when large wave numbers or highly oscillatory solutions are involved @xcite . for the helmholtz problem ( [ pde])-([bc ] ) , the corresponding variational form is given by seeking @xmath7 satisfying @xmath8 where @xmath9 and @xmath10 . in a classic finite element procedure ,
continuous polynomials are used to approximate the true solution @xmath11 . in many situations ,
the use of discontinuous functions in the finite element approximation often provides the methods with much needed flexibility to handle more complicated practical problems .
however , for discontinuous polynomials , the strong gradient @xmath12 in ( [ wf ] ) is no longer meaningful .
recently developed weak galerkin finite element methods @xcite provide means to solve this difficulty by replacing the differential operators by the weak forms as distributions for discontinuous approximating functions .
weak galerkin ( wg ) methods refer to general finite element techniques for partial differential equations and were first introduced and analyzed in @xcite for second order elliptic equations . through rigorous error analysis , optimal order of convergence of the wg solution in both discrete @xmath13 norm and @xmath14 norm
is established under minimum regularity assumptions in @xcite .
the mixed weak galerkin finite element method is studied in @xcite .
the wg methods are by design using discontinuous approximating functions . in this paper
, we will apply wg finite element methods @xcite to the helmholtz equation . the wg finite element approximation to ( [ wf ] )
can be derived naturally by simply replacing the differential operator gradient @xmath12 in ( [ wf ] ) by a weak gradient @xmath15 : find @xmath16 such that for all @xmath17 we have @xmath18 where @xmath19 and @xmath20 represent the values of @xmath21 in the interior and the boundary of the triangle respectively .
the weak gradient @xmath15 will be defined precisely in the next section .
we note that the weak galerkin finite element formulation ( [ wg1 ] ) is simple , symmetric and parameter free . to fully explore the potential of the wg finite element formulation ( [ wg1 ] )
, we will investigate its performance for solving the helmholtz problems with large wave numbers .
it is well known that the numerical performance of any finite element solution to the helmholtz equation depends significantly on the wave number @xmath22 .
when @xmath22 is very large representing a highly oscillatory wave , the mesh size @xmath23 has to be sufficiently small for the scheme to resolve the oscillations . to keep a fixed grid resolution , a natural rule is to choose @xmath24 to be a constant in the mesh refinement , as the wave number @xmath22 increases @xcite
however , it is known @xcite that , even under such a mesh refinement , the errors of continuous galerkin finite element solutions deteriorate rapidly when @xmath22 becomes larger .
this non - robust behavior with respect to @xmath22 is known as the `` pollution effect '' . to the end of alleviating the pollution effect ,
various continuous or discontinuous finite element methods have been developed in the literature for solving the helmholtz equation with large wave numbers @xcite .
a commonly used strategy in these effective finite element methods is to include some analytical knowledge of the helmholtz equation , such as characteristics of traveling plane wave solutions , asymptotic solutions or fundamental solutions , into the finite element space . likewise
, analytical information has been incorporated in the basis functions of the boundary element methods to address the high frequency problems @xcite . on the other hand , many spectral methods , such as local spectral methods @xcite , spectral galerkin methods @xcite , and spectral element methods @xcite have also been developed for solving the helmholtz equation with large wave numbers .
pollution effect can be effectively controlled in these spectral type collocation or galerkin formulations , because the pollution error is directly related to the dispersion error , i.e. , the phase difference between the numerical and exact waves @xcite , while the spectral methods typically produce negligible dispersive errors .
the objective of the present paper is twofold .
first , we will introduce weak galerkin methods for the helmholtz equation .
the second aim of the paper is to investigate the performance of the wg methods for solving the helmholtz equation with high wave numbers . to demonstrate the potential of the wg finite element methods in solving high frequency problems
, we will not attempt to build the analytical knowledge into the wg formulation ( [ wg1 ] ) and we will restrict ourselves to low order wg elements .
we will investigate the robustness and effectiveness of such plain wg methods through many carefully designed numerical experiments .
the rest of this paper is organized as follows . in section 2
, we will introduce a weak galerkin finite element formulation for the helmholtz equation by following the idea presented in @xcite .
implementation of the wg method for the problem ( [ pde])-([bc ] ) is discussed in section 3 . in section 4
, we shall present some numerical results obtained from the weak galerkin method with various orders .
finally , this paper ends with some concluding remarks .
let @xmath25 be a partition of the domain @xmath5 with mesh size @xmath23 .
assume that the partition @xmath25 is shape regular so that the routine inverse inequality in the finite element analysis holds true ( see @xcite ) .
denote by @xmath26 the set of polynomials in @xmath27 with degree no more than @xmath22 , and @xmath28 , @xmath29 , the set of polynomials on each segment ( edge or face ) of @xmath30 with degree no more than @xmath22 . for @xmath31 and given @xmath32
, we define the weak galerkin ( wg ) finite element space as follows @xmath33 where @xmath34 and @xmath35 are the values of @xmath36 restricted on the interior of element @xmath27 and the boundary of element @xmath27 respectively .
since @xmath35 may not necessarily be related to the trace of @xmath34 on @xmath30 , we write @xmath37 . for a given @xmath38 , we define another vector space @xmath39 where @xmath40 is the set of homogeneous polynomials of degree @xmath22 and @xmath41 ( see @xcite ) .
we will find a locally defined discrete weak gradient from this space on each element @xmath27 .
the main idea of the weak galerkin method is to introduce weak derivatives for discontinuous functions and to use them in discretizing the corresponding variational forms such as ( [ wf ] ) .
the differential operator used in ( [ wf ] ) is a gradient .
a weak gradient has been defined in @xcite .
now we define approximations of the weak gradient as follows . for each @xmath42
, we define a discrete weak gradient @xmath43 on each element @xmath27 such that @xmath44 where @xmath45 is locally defined on each element @xmath27 , @xmath46 and @xmath47 .
we will use @xmath48 to denote @xmath49 .
then the wg method for the helmholtz equation ( [ pde])-([bc ] ) can be stated as follows .
a numerical approximation for ( [ pde ] ) and ( [ bc ] ) can be obtained by seeking @xmath50 such that for all @xmath51 @xmath52 denote by @xmath53 the @xmath14 projection onto @xmath54 , @xmath55 . in other words , on each element @xmath27 , the function @xmath56 is defined as the @xmath14 projection of @xmath11 in @xmath26 and @xmath57 is the @xmath14 projection of @xmath11 in @xmath58 . for equation ( [ pde ] ) with dirichlet boundary condition @xmath59 on @xmath60 ,
optimal error estimates have been obtained in @xcite . for a sufficiently small mesh size @xmath23
, we can derive following optimal error estimate for the helmholtz equation ( [ pde ] ) with the mixed boundary condition ( [ bc ] ) .
let @xmath16 and @xmath61 be the solutions of ( [ wg ] ) and ( [ pde])-([bc ] ) respectively and assume that @xmath5 is convex .
then for @xmath31 , there exists a constant @xmath62 such that @xmath63 the proof of this theorem is similar to that of theorem 8.3 and theorem 8.4 in @xcite and is very long .
since the emphasis of this paper is to investigate the performance of the wg method , we will omit details of the proof .
first , define a bilinear form @xmath64 as @xmath65 then ( [ wg ] ) can be rewritten with @xmath66 @xmath67 the methodology of implementing the wg methods is the same as that for continuous galerkin finite element methods except that the standard gradient operator @xmath12 should be replaced by the discrete weak gradient operator @xmath15 . in the following , we will use the lowest order weak galerkin element ( @xmath22=0 ) on triangles as an example to demonstrate how one might implement the weak galerkin finite element method for solving the helmholtz problem ( [ pde ] ) and ( [ bc ] ) .
let @xmath68 and @xmath69 denote , respectively , the number of triangles and the number of edges associated with a triangulation @xmath25 .
let @xmath70 denote the union of the boundaries of the triangles @xmath27 of @xmath25 .
the procedure of implementing the wg method ( [ wg ] ) consists of the following three steps . 1 .
find basis functions for @xmath71 defined in ( [ vh ] ) : @xmath72 where @xmath73 and @xmath74 0 \quad \mbox{otherwise } , \end{array } \right .
\psi_j=\left\ { \begin{array}{l } 1 \quad \mbox{on } \;\
; e_j , \\ [ 0.08 in ] 0 \quad \mbox{otherwise } , \end{array } \right.\ ] ] for @xmath75 and @xmath76 .
please note that @xmath77 and @xmath78 are defined on whole @xmath5 .
+ 2 . substituting @xmath79 into ( [ wg8 ] ) and letting @xmath80 in ( [ wg8 ] )
yield @xmath81 where @xmath82 and @xmath83 are the values of @xmath84 on the interior of the triangle and the boundary of the triangle respectively . in our computations ,
the integrations on the right - hand side of ( [ sys ] ) are conducted numerically . in particular , a 7-points two - dimensional gaussian quadrature and a 3-points one - dimensional gaussian quadrature are employed , respectively , to calculate @xmath85 and @xmath86 numerically .
form the coefficient matrix @xmath87 of the linear system ( [ sys ] ) by computing @xmath88 all integrations in ( [ bilinear ] ) are carried out analytically .
finally , we will explain how to compute the weak gradient @xmath15 for a given function @xmath89 when @xmath90 .
for a given @xmath38 , we will find @xmath91 , @xmath92 for example , we can choose @xmath93 as follows @xmath94 thus on each element @xmath95 , @xmath96 . using the definition of the discrete weak gradient ( [ d - g ] ) ,
we find @xmath97 by solving the following linear system : @xmath98 the inverse of the above coefficient matrix can be obtained explicitly or numerically through a local matrix solver . for the basis function @xmath84 ,
@xmath99 is nonzero on only one or two triangles .
in this section , we examine the wg method by testing its accuracy , convergence , and robustness for solving two dimensional helmholtz equations . the pollution effect due to
large wave numbers will be particularly investigated and tested numerically . for convergence tests ,
both piecewise constant and piecewise linear finite elements will be considered . to demonstrate the robustness of the wg method , the helmholtz equation in both homogeneous and inhomogeneous media will be solved on convex and non - convex computational domains . the mesh generation and all computations
are conducted in the matlab environment . for simplicity ,
a structured triangular mesh is employed in all cases , even though the wg method is known to be very flexible in dealing with various different finite element partitions @xcite .
two types of relative errors are measured in our numerical experiments .
the first one is the relative @xmath14 error defined by @xmath100 the second one is the relative @xmath13 error defined in terms of the discrete gradient @xmath101 numerically , the @xmath13-semi - norm will be calculated as @xmath102 for the lowest order finite element ( i.e. , piecewise constants ) . for piecewise linear elements , we use the original definition of @xmath15 to compute the @xmath13-semi - norm @xmath103 .
cc we first consider a homogeneous helmholtz equation defined on a convex hexagon domain , which has been studied in @xcite .
the domain @xmath5 is the unit regular hexagon domain centered at the origin @xmath104 , see fig .
[ fig.domain ] ( left ) .
here we set @xmath105 and @xmath106 in ( [ pde ] ) , where @xmath107 . the boundary data @xmath108 in the robin boundary condition ( [ bc ] )
is chosen so that the exact solution is given by @xmath109 where @xmath110 are bessel functions of the first kind .
let @xmath32 denote the regular triangulation that consists of @xmath111 triangles of size @xmath112 , as shown in fig .
[ fig.domain ] ( left ) for @xmath113 .
.convergence of piecewise constant wg for the helmholtz equation on a convex domain with wave number @xmath114 .
[ cols= " < , < , < ,
< , < " , ] we finally investigate the performance of the wg method for the helmholtz equation with large wave numbers . as discussed above , without resorting to high order generalizations or analytical / special treatments , we will examine the use of the plain wg method for tackling the pollution effect .
the homogeneous helmholtz problem of the subsection [ convex ] will be studied again .
also , the @xmath115 and @xmath116 elements are used to solve the homogeneous helmholtz equation with the robin boundary condition . since this problem is defined on a structured hexagon domain , a uniform triangular mesh with a constant mesh size @xmath23 throughout the domain is used .
this enables us to precisely evaluate the impact of the mesh refinements .
following the literature works @xcite , we will focus only on the relative @xmath13 semi - norm in the present study .
cc to study the non - robustness behavior with respect to the wave number @xmath22 , i.e. , the pollution effect , we solve the corresponding helmholtz equation by using piecewise constant wg method with various mesh sizes for four wave numbers @xmath117 , @xmath118 , @xmath119 , and @xmath120 , see fig .
[ fig.kh ] ( left ) for the wg performance . from fig .
[ fig.kh ] ( left ) , it can be seen that when @xmath23 is smaller , the wg method immediately begins to converge for the cases @xmath117 and @xmath118 .
however , for large wave numbers @xmath119 and @xmath120 , the relative error remains to be about @xmath121% , until @xmath23 becomes to be quite small or @xmath122 is large .
this indicates the presence of the pollution effect which is inevitable in any finite element method @xcite . in the same figure
, we also show the errors of different @xmath22 values by fixing @xmath123 .
surprisingly , we found that the relative @xmath13 error does not evidently increase as @xmath22 becomes larger .
the convergence line for @xmath123 looks almost flat , with a very little slope .
in other words , the pollution error is very small in the present wg result .
we note that such a result is as good as the one reported in @xcite by using a penalized discontinuous galerkin approach with optimized parameter values .
in contrast , no parameters are involved in the wg scheme . on the other hand
, the good performance of the wg method for the case @xmath123 does not mean that the wg method could be free of pollution effect .
in fact , it is known theoretically @xcite that the pollution error can not be eliminated completely in two- and higher - dimensional spaces for galerkin finite element methods . in the right chart of fig .
[ fig.kh ] , we examine the numerical errors by increasing @xmath22 , under the constraint that @xmath24 is a constant . huge wave numbers , up to @xmath124 , are tested .
it can be seen that when the constant changes from @xmath125 to @xmath126 and @xmath127 , the non - robustness behavior against @xmath22 becomes more and more evident .
however , the slopes of @xmath24=constant lines remain to be small and the increment pattern with respect to @xmath22 is always monotonic .
this suggests that the pollution error is well controlled in the wg solution .
cc c + + cc c + in the rest of the paper , we shall present some numerical results for the wg method when applied to a challenging case of high wave numbers . in fig .
[ fig.solu2d ] and [ fig.solu2d_linear ] , the wg numerical solutions are plotted against the exact solution of the helmholtz problem . here
we take a wave number @xmath120 and mesh size @xmath128 which is relatively a coarse mesh .
with such a coarse mesh , the wg method can still capture the fast oscillation of the solution .
however , the numerically predicted magnitude of the oscillation is slightly damped for waves away from the center when piecewise constant elements are employed in the wg method .
such damping can be seen in a trace plot along @xmath129-axis or @xmath130 . to see this
, we consider an even worse case with @xmath120 and @xmath131 .
the result is shown in the first chart of fig .
[ fig.solu1d ] .
we note that the numerical solution is excellent around the center of the region , but it gets worse as one moves closer to the boundary .
if we choose a smaller mesh size @xmath132 , the visual difference between the exact and wg solutions becomes very small , as illustrate in fig .
[ fig.solu1d ] .
if we further choose a mesh size @xmath133 , the exact solution and the wg approximation look very close to each other .
this indicates an excellent convergence of the wg method when the mesh is refined .
in addition to mesh refinement , one may also obtain a fast convergence by using high order elements in the wg method .
figure [ fig.solu1d_linear ] illustrates a trace plot for the case of @xmath120 and @xmath128 when piecewise linear elements are employed in the wg method .
it can be seen that the computational result with this relatively coarse mesh captures both the fast oscillation and the magnitude of the exact solution very well .
the present numerical experiments indicate that the wg method as introduced in @xcite is a very promising numerical technique for solving the helmholtz equations with large wave numbers .
this finite element method is robust , efficient , and easy to implement . on the other hand , a theoretical investigation for the wg method
should be conducted by taking into account some useful features of the helmholtz equation when special test functions are used .
it would also be valuable to test the performance of the wg method when high order finite elements are employed to the helmholtz equations with large wave numbers in two and three dimensional spaces . finally , it is appropriate to clarify some differences and connections between the wg method and other discontinuous finite element methods for solving the helmholtz equation .
discontinuous functions are used to approximate the helmholtz equation in many other finite element methods such as discontinuous galerkin ( dg ) methods @xcite and hybrid discontinuous galerkin ( hdg ) methods @xcite
. however , the wg method and the hdg method are fundamentally different in concept and formulation .
the hdg method is formulated by using the standard mixed method approach for the usual system of first order equations , while the key to the wg is the use of the discrete weak differential operators . for a second order elliptic
problem , these two methods share the same feature by approximating first order derivatives or fluxes through a formula that was commonly employed in the mixed finite element method . for high order partial differential equations ( pdes ) ,
the wg method is greatly different from the hdg .
consider the biharmonic equation @xcite as an example .
the first step of the hdg formulation is to rewrite the fourth order equation to four first order equations . in contrast
, the wg formulation for the biharmonic equation can be derived directly from the variational form of the biharmonic equation by replacing the laplacian operator @xmath134 by a weak laplacian @xmath135 and adding a parameter free stabilizer @xcite .
it should be emphasized that the concept of weak derivatives makes the wg a widely applicable numerical technique for a large variety of pdes which we shall report in forthcoming papers . for the helmholtz equation studied in this paper , the wg method and
the hdg method yield the same variational form for the homogeneous helmholtz equation with a constant @xmath136 in ( [ pde ] ) .
however , the wg discretization differs from the hdg discretization for an inhomogeneous media problem with @xmath136 being a spatial function of @xmath129 and @xmath137 .
moreover , the wg method has an advantage over the hdg method when the coefficient @xmath136 is degenerated .
i. babuka , f. ihlenburg , e.t .
paik , s.a .
sauter , a generalized finite element method for solving the helmholtz equation in two dimensions with minimal pollution .
_ computer methods in applied mechanics and engineering _ 1995 ; * 128 * : 325 - 359 .
i. babuka , s.a .
sauter , is the pollution effect of the fem avoidable for the helmholtz equation considering high wave number ?
_ siam journal on numerical analysis _ 1997 ; * 34 * : 2392 - 2423 . reprinted in _ siam review _ 2000 ; * 42 * : 451 - 484 . f. ihlenburg , i. babuka , dispersion analysis and error estimation of galerkin finite element methods for the helmholtz equation .
_ international journal for numerical methods in engineering _ 1995 ; * 38 * : 3745 - 3774 .
f. ihlenburg , i. babuka , finite element solution of the helmholtz equation with high wavenumber part ii : the @xmath23-@xmath138-version of the fem .
_ siam journal of numerical analysis _ 1997 ; * 34 * : 315 - 358 . | a weak galerkin ( wg ) method is introduced and numerically tested for the helmholtz equation . this method is flexible by using discontinuous piecewise polynomials and retains the mass conservation property . at the same time , the wg finite element formulation is symmetric and parameter free .
several test scenarios are designed for a numerical investigation on the accuracy , convergence , and robustness of the wg method in both inhomogeneous and homogeneous media over convex and non - convex domains . challenging problems with high wave numbers
are also examined .
our numerical experiments indicate that the weak galerkin is a finite element technique that is easy to implement , and provides very accurate and robust numerical solutions for the helmholtz problem with high wave numbers .
galerkin finite element methods , discrete gradient , helmholtz equation , large wave numbers , weak galerkin primary , 65n15 , 65n30 , 76d07 ; secondary , 35b45 , 35j50 |
lyman - break galaxies have given us our first detailed look at star formation at high redshifts ( @xmath10 ) .
however , our picture of the distant universe is limited because these objects are selected for strong rest - frame ultraviolet emission arising from substantial unobscured starbursts in the galaxies .
individually , they tell us little about darker baryons at the same redshift , whether the systems containing the baryons are intrinsically less uv - luminous or are obscured by dust . the lyman - break galaxy ( lbg ) population
has now been probed from @xmath11 @xcite up to @xmath12 @xcite , with tentative detections reported at redshifts as high as @xmath13 @xcite .
the photometric and spectroscopic properties of lbgs evolve with increasing redshift .
typical lbgs at @xmath14 have stellar masses of a few times 10@xmath15m@xmath3 , are predominantly young ( @xmath16myr ) , and are sub - solar but not primordial in their metallicity ( [email protected]@xmath3 ) , where these properties are deduced from their rest - frame optical and near - infrared spectral energy distribution @xcite .
however something is missing in our picture of the universe at @xmath18 .
the youth of lyman - break objects when compared to the range of lookback times probed in a given survey suggests that only a small fraction of the overall population is undergoing a phase of strong unobscured star formation , and hence is detectable in the rest uv , at a given time .
fossil stellar populations in the most massive galaxies at the present time suggest that the bulk of their stars were formed at @xmath19 @xcite , but the mass assembly history of the universe as traced by lbgs implies that only 1% of the baryonic material that will eventually form today s galaxies was in collapsed systems at @xmath20 .
where is the other 99% ?
the bulk of the baryonic material at these high redshifts must be in either low mass systems or the gas phase .
damped lyman - alpha systems in the spectra of distant quasars have shown that dense neutral gas is ubiquitous at high redshift , holding a third of the hi atoms at @xmath20 and appearing in 10% of sight - lines per unit cosmological distance @xcite .
these sources represent reservoirs of neutral gas for star formation , yet are known to be deficient in h@xmath1 relative to local galaxies .
molecular gas is instead more tightly concentrated in high - density , relatively dusty regions @xcite .
securely identifying this material is challenging . at @xmath14 ,
the far - infrared lines emitted by cool and obscured material are redshifted to millimeter wavelengths , but not necessarily into an atmospheric window . at these extreme luminosity distances lines
are faint , requiring long integration times to secure even a tentative detection .
the fields of view of existing millimeter interferometers are small , and their correlators have , until recently , been limited to bandwidths of a few hundred mhz ( equivalent to @xmath21 ) .
the resulting small volumes made survey work at high redshift unfeasible . as a result
, the few detections of line emission from molecular and atomic gas at high redshift have been towards active galactic nuclei - quasars to @xmath22 , radio galaxies to @xmath23 @xcite - and hence probed highly atypical environments . in this paper
we present the results of a pilot programme to explore the cool gas associated with a large scale structure at @xmath18 marked out by lyman - break galaxies rather than by an active galaxy .
our 40 arcmin@xmath24 target field contains seven uv - luminous sources with spectroscopically confirmed redshifts in a narrow range ( @xmath25 ) , a 6@xmath5 excess over the typical density of such sources in our survey , but shows no clear spatial clustering .
given the redshift dispersion and spatial distribution of these galaxies , it is clear that their star formation can not be a response to a common triggering event , and yet the probability of such a configuration arising by chance is small .
the simplest explanation is that there is a large mass of hidden baryons in this field at this redshift . at any one time
this structure is traced by the small fraction of that material undergoing comparatively short - term uv - luminous star formation events and appearing as lbgs .
while the stars within the lbgs are likely to end up in the most massive current - day galaxies after subsequent mergers , the total stellar mass in even this overdensity of lbgs is an order of magnitude too small to account for the stars in the most massive spheroids . consequently , there should be far more baryons in the vicinity of the lbgs than revealed in the uv .
our pilot programme was designed to probe the cool gas mass in a small section of the structure , centered on a pair of lbgs with lyman-@xmath26 emission redshifts separated by @xmath27 .
the observations target not only the two lbgs , but any darker / fainter systems in the underlying large scale structure .
the results for lower redshift lbgs ( e.g. * ? ? ?
* ) , suggested that we were unlikely to detect the two known galaxies directly .
all magnitudes in this paper are quoted in the ab system @xcite .
we adopt a @xmath28cdm cosmology with ( @xmath29 , @xmath30 , @xmath31)=(0.7 , 0.3 , 0.7 ) .
the observations were carried out at the australia telescope compact array in march - april 2008 .
the array was in the relatively compact h168 configuration , with both north - south and east - west baselines .
we used the correlator in the 128 mhz/64 channel configuration with two overlapping intermediate frequency ( if ) bands , giving a total bandwidth of 240mhz , and measuring the total incident flux density , but not polarization .
we investigated two transitions of co at high redshift . in order to constrain the co(1 - 0 ) line
, we tuned the if bands to 18.838ghz and 18.958ghz ( allowing detection of an emission line in the range @xmath32 ) . in this configuration ,
the spectral resolution is 4mhz or 63.5kms@xmath8 .
observations were taken in three 8hr observing periods on the nights of 2008 march 22 - 24 ( although data taken on the 23rd were discarded due to poor weather ) . a single pointing was observed for the entire observing period , and a nearby bright source ( 1045 - 188 ) observed every 15 mins to determine phase stability . pointing accuracy was checked every hour .
absolute flux calibration was through observations of 1921 - 293 ( the standard flux calibrator at the atca ) each night .
the half - power - beam - width ( hpbw ) of the atca at this frequency is 2.5 arcmin and the naturally - weighted restoring beam in this configuration is 15.9@xmath33 10.9 .
the co(2 - 1 ) line was observed in four 11hr observations on the nights of 2008 apr 04 - 07 using the recently - commissioned 7 mm receiver system .
the if bands were tuned to 37.672ghz and 37.792ghz ( giving redshift coverage @xmath34 ) .
a single pointing centered on the 18.8ghz field center was observed , and 1045 - 188 was used as a pointing and phase calibrator .
uranus , the primary flux calibrator for the atca at millimeter wavelengths , was not visible at any point in our observing period .
instead we bootstrap our primary flux calibration from two compact hii regions , observed each night , which were also observed using our configuration by atca staff on the afternoon of 2008 apr 04 together with uranus and a bandpass calibrator . at this frequency , the hpbw is 74 and the restoring beam is 7.3@xmath33 4.8 .
the total useful observing time was 16 hours at 18.9ghz and 44 hours at 37.7ghz .
the rms noise in the final images is 0.11 mjy / beam and 0.10 mjy / beam in each 4mhz channel at 18.9 and 37.7ghz respectively . [
cols="^,<,^,^,^,^ " , ] [ tab : sources ]
as illustrated in figure [ fig : image ] , our pointing included two lyman-@xmath26 emitting , lyman - break galaxies within the hpbw at 18.9ghz ( table [ tab : sources ] ) . at 37.7ghz
one of these ( at @xmath35 ) lies well outside the beam half - power point , while the second ( at @xmath36 ) remains close to the pointing center .
neither galaxy shows evidence for line or continuum emission in either of the two transitions surveyed .
this is not , in itself , surprising given the youth and low masses ( typically a few @xmath37m@xmath3 , * ? ? ?
* ) of @xmath20 lyman - break galaxies . however , these observations provide the strongest constraint yet on the molecular gas content of @xmath18 star - forming galaxies .
we use our limiting flux at the frequency of the redshifted co(1 - 0 ) line to constrain the molecular gas content of these galaxies .
we employ the widely - used conversion factor of m@xmath38/@xmath39=0.8 m@xmath3 ( kkms@xmath8pc@xmath24)@xmath8 based on local infrared - luminous galaxies .
our image rms noise of 0.11 mjy / beam at the co(1 - 0 ) line corresponds to a formal 2 sigma limit @xmath40 kkms@xmath8pc@xmath41 for the luminosity of an unresolved 200 km s@xmath8 emission line at @xmath0 .
thus we constrain the molecular gas mass of the known lyman - break galaxies in this field to be m(h@xmath1)@xmath2m@xmath3 ( 2@xmath5 ) .
one of our lbgs ( source b in table [ tab : sources ] ) is also close to the center of the beam in our 37ghz observations . for thermalized , optically - thick co emission ,
the line luminosity is independent of transition . making this assumption , non - detection of the 2 - 1
line places a tighter constraint on the mass of m(h@xmath1)@xmath42m@xmath3 ( 2@xmath5 ) , comparable to the inferred stellar mass in such sources .
inspection of the atca data cubes revealed no evidence for continuum sources .
only one emission feature was found to extend over more than two adjacent frequency channels .
this feature , a spatially - unresolved emission line centered at 37.642ghz with a fwhm of 110kms@xmath8 ( equivalent to four independent frequency channels ) , lies on the beam half - power point .
it has a peak flux of @xmath43mjy / beam , and a velocity integrated flux of @xmath7 jy km s@xmath8 . a combined frame of channels not containing line emission ( a total bandwidth of 180mhz ) shows no significant continuum emission with a formal limit on the continuum of @xmath440.49 mjy / beam at 37.73ghz .
the line emission is centered 1 arcminute ( 370kpc , comoving ) north - west and 40 arcseconds ( 250kpc ) north - east of our spectroscopically confirmed lyman - break galaxies at @xmath36 and @xmath35 respectively , towards more lbgs in the field .
this line emitter is undetected at 18.9ghz and not spatially coincident with a known source at any observed wavelength ( from 0.5@xmath45 m to 21 cm ) .
our optical and ir data require that the host galaxy must have a flux density @xmath46jy at an observed frame wavelength of 4.5@xmath45 m ( 0.7@xmath45 m rest if at @xmath47 ) and @xmath48jy at a 0.8@xmath45 m ( 1300 rest ) .
no sub - mm data exist at this location .
we note that the lensing properties of this field are well understood , and we use the lensing maps of to estimate a maximum 10% enhancement in the flux of a source at this location . in order to verify the reality of the 37.6ghz line , we investigated its properties thoroughly .
the velocity - integrated line is a formal 8@xmath5 detection in a multi - frequency synthesis image of 6 channels centered on the line emission .
it is detected in the @xmath49-dataset on multiple baselines , and in two independent sub - sets of the data .
its strength in the flux - calibrated image is independent of image deconvolution algorithm and the location of the imaging center .
we conclude that this is unlikely to be a spurious detection .
we evaluate the probability that this is a line emitter at a redshift unrelated to our high redshift system ( i.e. not a co(2 - 1 ) emission line ) , by considering the population models of @xcite .
these authors calculated the expected number density of line emitters as a function of frequency and line flux , given models for galaxy and dark matter distributions at @xmath50 . as a cautious assumption
, we consider the source counts predicted at 50ghz at our integrated flux density ( while noting that source counts at 37.6ghz are likely to be lower ) . in our 256mhz total bandwidth , and allowing for source detections within 1.5 hpbw , we would expect 0.003 line emitters in our survey , with these dominated by redshifted co(1 - 0 ) lines at @xmath51 .
coupled with the known existence of structure at @xmath0 in this field , we interpret the line detection as co(2 - 1 ) emission at @xmath6 at high confidence . at the measured line flux in the co(2 - 1 ) transition , and assuming constant luminosity , the expected counterpart in the 1 - 0 transition would be below the 2@xmath5 level at 18.8ghz , consistent with our non - detection .
using the same co - to - h@xmath1 conversion factor discussed in section [ sec : lbg_limits ] applied to the luminosity of the co(2 - 1 ) transition , we infer a molecular gas mass of m(h@xmath1)@xmath52 m@xmath3 in this source .
the beam width at 37.7ghz implies an upper limit on the physical size of our unresolved source of 45@xmath3330kpc . a simple virial calculation assuming a uniform mass distribution suggests that our derived gas mass could represent up to 50% of the dynamical mass in the system , while remaining both bound and smaller than our beam size .
this pilot survey at 37.7ghz probed only 3% of the projected spatial distribution and 35% of the redshift range occupied by clustered uv - luminous galaxies in our target field . despite this , the gas mass inferred from our single detection is comparable to the total stellar mass in uv - luminous sources in the structure .
if this detection is typical of the space density of molecular line emitters in this field , then galaxies with unobscured star formation represent at best a few percent of the baryons present ( consistent with * ? ? ?
* ) . in the nomenclature of
, our detected source is an early universe molecular emission line galaxy ( emg ) , a broad category including radio galaxies , sub - mm galaxies and qsos at @xmath53 .
it represents @xmath54m@xmath3 of co , formed within 1.1gyr of the big bang .
this yield is plausible from @xmath55 population iii stars , each of 100m@xmath3 , in as little as the stellar lifetime ( 10myr , * ? ? ?
by contrast , the yield of more conventional supernovae and the typical star formation rates of uv - luminous sources in this field ( @xmath5630m@xmath3yr@xmath8 ) , @xmath57 years would be required ( see * ? ? ? * ) implying either early and continuous star formation , or an early , intense starburst . a source with unobscured star formation of 10m@xmath3yr@xmath8 at @xmath0
would have been detected in our deep optical imaging , but observations of dust emission in the rest - frame far - infrared will be necessary to distinguish between a mature stellar population at the time of observation and ongoing , but dust - obscured , star formation .
it is interesting to compare our detection with a similar observation at @xmath23 , also using the atca .
@xcite studied a high mass system at this redshift , marked out by the presence of at least six lyman-@xmath26 emitter galaxies and a powerful radio galaxy . based on uv - luminous stellar mass
, this system should be comparable in mass to our target region .
@xcite detected co emission in the 1 - 0 and 5 - 4 transitions , clearly associated with the radio galaxy . while the inferred gas mass is within a factor of four of our object , the velocity structure differs : the emission line detected by @xcite was twice as broad as our line emitter , possibly due to momentum input from the radio source .
we note that the co luminosity and derived mass of our line - emitter are similar to those of sub - millimeter galaxies ( smgs ) at lower redshifts @xcite .
these are also often a ssociated with galaxy overdensities at high redshift ( e.g. * ? ? ?
* ) . while some smgs with this mass would have been detected in our deep
spitzer_/irac imaging ( based on their @xmath58-band magnitudes at @xmath59 , * ? ? ?
* redshifted to @xmath0 ) , others would not .
co emission is not a reliable star formation rate ( sfr ) indicator , but is loosely correlated with far - infrared luminosity , and hence sfr .
data on this relation at high redshift is sparse , but known emgs with comparable gas masses to our detected source have sfrs of @xmath60m@xmath3yr@xmath8 .
by contrast , the tight limit on h@xmath1 mass in the lbgs in our field would imply obscured star formation rates of less than a few hundred solar masses per year , even given dust properties similar to those of smgs .
our detected source likely represents a more extreme environment than that seen in lbgs at the same redshift and could be a high redshift counterpart for the intense starbursts in smgs .
it is becoming increasingly clear that no one survey technique can provide a complete picture of large scale structure at high redshift .
overdensities of lyman break galaxies are correlated with the presence of lyman - alpha emitters , radio galaxies , smgs , dlas @xcite , and now molecular emission line galaxies at the same redshift .
each component of the structure may account for only a few percent of the baryonic mass and each presents observational challenges . while lbgs and dlas probe comparable mass haloes at @xmath61 @xcite
, molecular line emitters may probe more massive and denser regions of the same structures .
extensive extrapolation based on a single line from a single source is risky .
observations of this source in other emission lines will be required to confirm its redshift and luminosity .
such observations will also begin to characterize the gas temperature and density , and determine whether the co to gas mass conversion factor for ulirgs is appropriate for low mass systems at early times .
additional sources must be identified and characterized to determine whether our initial detection is anomalous or typical of its population .
ultimately , more extensive millimeter surveys , based on existing deep datasets , are required to establish the molecular gas content of large scale structures at high redshift , and to probe this early stage of galaxy formation .
our main conclusions can be summarized as follows : \(ii ) neither of two known lyman break galaxies in our survey area show associated molecular line emission , suggesting that their molecular gas content does not significantly exceed their stellar content .
ers acknowledges support from the uk stfc .
based on data from atca programs c1753 and c1821 .
the australia telescope compact array is part of the australia telescope which is funded by the commonwealth of australia for operation as a national facility managed by csiro .
the authors thank chris carilli , steve longmore and ron ekers for useful discussions .
we also thank maxim voronkov and robin ward for their vital assistance as duty astronomers at the atca . | we present observations of redshifted co(1 - 0 ) and co(2 - 1 ) in a field containing an overdensity of lyman break galaxies ( lbgs ) at @xmath0 .
our australia telescope compact array observations were centered between two spectroscopically - confirmed @xmath0 galaxies .
we place upper limits on the molecular gas masses in these two galaxies of m(h@xmath1)@xmath2 m@xmath3 and @xmath4 m@xmath3 ( 2@xmath5 ) , comparable to their stellar masses .
we detect an optically - faint line emitter situated between the two lbgs which we identify as warm molecular gas at @xmath6 .
this source , detected in the co(2 - 1 ) transition but undetected in co(1 - 0 ) , has an integrated line flux of @xmath7 jy kms@xmath8 , yielding an inferred gas mass m(h@xmath1)=@xmath9m@xmath3 .
molecular line emitters without detectable counterparts at optical and infrared wavelengths may be crucial tracers of structure and mass at high redshift . |
with the discovery of the cosmological nature of @xmath0-ray bursts ( grbs ) in 1997 @xcite , astronomers were suddenly forced to explain the enormous isotropic energy release of these distant explosions .
some of the most energetic events , such as grb990123 , seemingly released enough energy in the prompt @xmath0-rays ( @xmath4 ergs ; @xcite ) to rival the rest mass of a neutron star .
furthermore , broadband modeling of the best sampled events has shown that a comparable amount of energy remains in the shock , powering the long - lived x - ray , optical , and radio afterglow ( see e.g. @xcite ) .
the hypothesis that grbs are a - spherical explosions @xcite , supported by the appearance of achromatic `` jet '' breaks in a large number of afterglow light curves @xcite , proved to be a turning point . with typical opening angles of a few degrees ,
the true energy release from most grbs is @xmath5 ergs , on par with that of a supernova ( sn ) .
this realization enabled the discovery of a standard energy reservoir for the collimation - corrected prompt energy @xcite and kinetic energy of the afterglow @xcite .
grbs are now considered promising standard candle candidates , with the hope of hubble diagrams out to @xmath6 offering complementary constraints to type ia sne on the cosmology of our universe ( @xcite ; c.f .
. launched in november 2004 , the _ swift _ gamma - ray burst explorer @xcite was designed to position grbs , disseminate accurate coordinates to ground - based observatories in real - time , and follow the uv and x - ray afterglows from minutes to days after the event . in only a year of full operation , _ swift _ has brought about a number of fundamental advances in the grb field , including the discovery of the first x - ray ( grb050509b ; @xcite ) and near - infrared ( grb050724 ; @xcite ) afterglows of a short - hard burst , the detection of the high - redshift ( @xmath7 ) grb050904 @xcite , and the ability to measure broadband light curves starting shortly after , and in a few cases even during the @xmath0-ray event itself . despite these advances , measuring the bolometric fluences of _ swift _ events
has proved challenging , for a number of reasons .
first , the limited energy range of the _ swift _ burst alert telescope ( bat ; @xcite ) means that _ swift _ can accurately characterize only the softest grb spectra .
second , few _ swift _ events have shown conclusive signs of a jet break , leaving geometric corrections highly uncertain .
finally , the x - ray light curves of _ swift _ afterglows have shown both bright flares @xcite and slow decays @xcite . both behaviors have been attributed to late - time ( @xmath8 ) energy injection , and at times have rivaled the energy release of the prompt @xmath0-ray emission ( see e.g. @xcite ) .
on 20 august 2005 ut , the bat detected and localized the unusually bright and long grb050820a , a truly rare burst in the _ swift _ sample .
the @xmath0-ray light curve is marked by a soft pulse of emission preceding the main event by over 200 s. the main emission was bright enough to be detected by the _ konus - wind _ instrument , providing a @xmath0-ray spectrum extending beyond 1 mev as well as continuous coverage over the entire @xmath9 600 s burst duration .
since _ swift _ triggered on the precursor , both space- and ground - based facilities were able to image the transient during the bulk of the prompt emission .
such contemporaneous multi - wavelength observations have only been achieved for a handful of bursts to date .
the bright x - ray ( @xmath10 mjy ) and optical ( @xmath11 mag ) afterglows made it possible to study the evolution of the afterglow for weeks after the burst , providing one of the most detailed broadband light curves in the _ swift _ era
. finally , late - time _ hubble space telescope _
observations allowed us to constrain the jet - break time , and hence the geometry of the outflow . even after applying the collimation correction
, we find grb050820a is an exceptionally energetic event . our work proceeds as follows : in [ sec : obs ] , we outline our broadband observations of the grb050820a , beginning with the high - energy prompt emission and followed by the x - ray , optical , and radio afterglow .
we find the afterglow data are incompatible with the standard model of synchrotron radiation from a single , highly - relativistic shock expanding adiabatically into the surrounding medium @xcite .
instead in [ sec : anal ] we use power - law fits ( @xmath12 ) to model the afterglow , dividing the burst into segments based on noticeable temporal breaks in the x - ray and optical light curves .
this analysis allows us to investigate the early broadband light curve ( [ sec : pre ] and [ sec : early ] ) , late - time ( @xmath13 ) energy injection in the forward shock ( [ sec : inject ] ) , the structure of the circum - burst medium ( [ sec : cbm ] ) , and the geometry and energetics of the event ( [ sec : energy ] ) . throughout this work
we adopt a standard cosmology with @xmath14 km s@xmath15 mpc@xmath15 , @xmath16 , and @xmath17 .
we also make use of the notation @xmath18 .
in this section , we present our broadband observations of grb050820a , which span the spectral range from @xmath0-rays to radio frequencies and extend in time out to 61 days after the burst .
we include an independent analysis of the _ swift _ bat data set , as well as the complete light curve and spectrum from the _ konus - wind _ instrument @xcite , which , unlike the bat , was able to observe grb050820a over its entire duration ( [ sec : gamma ] ) . in [ sec : xray ] we provide an analysis of the _ swift _ x - ray telescope ( xrt ; @xcite ) data , with afterglow detections out to two weeks after the event .
we present contemporaneous optical data from the automated palomar 60-inch telescope ( p60 ; cenko et al .
2006 , in preparation ) and the _ swift _ ultra - violet optical telescope ( uvot ; @xcite ) , supplemented by late - time images taken with the 9.2-m _ hobby - eberly telescope _ ( het ) and the _ hubble space telescope _ ( hst ) ( [ sec : opt ] ) . finally , we monitored grb050820a in the radio with the very large array beginning only hours after the burst and continuing for approximately two months ( [ sec : radio ] ) . at 06:34:53 on 20 august 2005 ut , for it is assumed to coincide with the beginning of the prompt emission .
given the unique nature of the high - energy emission from grb050820a , we undertake a more detailed study to determine exactly when the prompt emission began ( i.e. @xmath19 ) in [ sec : pre ] .
times measured with reference to the _ swift _ trigger time will be referred to as @xmath20 throughout the remainder of this work . ] , the bat triggered and located grb050820a ( _ swift _ trigger 151207 , @xcite ) .
the initial location calculated on - board was a 4 error circle centered at @xmath21 , @xmath22 ( j2000.0 ) .
@xcite describe a multi - peaked light curve ( @xmath23 s ) with clear spectral evolution ( hard - to - soft ) within each peak .
following the report of additional high - energy emission from the _ konus - wind _ instrument ( @xcite ; see [ sec : kw ] ) , the bat team re - analyzed their full light curve and found evidence of a much stronger , harder episode of emission from grb050820a @xcite .
unfortunately the satellite entered the south atlantic anomaly ( saa ) approximately 240 seconds after the burst trigger , and thus estimates of the properties of this second phase are highly uncertain .
here we have independently analyzed the bat data from grb050820a .
we have extracted the @xmath24 kev light curve in 1 s timing bins using software tools from the _ swift _ data analysis package .
the result is shown in figure [ fig : early_all]a .
in addition , we have extracted spectra for the two periods of high energy emission covered by the bat data ( peaks @xmath25 and @xmath26 , see [ sec : kw ] ) .
we then fit these spectra to a power - law distribution of energies ( @xmath27 ) .
we find evidence for strong spectral evolution between these two intervals , as the second peak is significantly harder than the first .
the results of this analysis are shown in table [ tab : bat ] .
rcc[b ] @xmath28 @xmath29 & @xmath30 & 1.07 / 75
+ @xmath31 @xmath32 & @xmath33 & 0.95 / 76 + [ tab : bat ] the main part of grb 050820a triggered _ konus - wind _ at @xmath34 = 06:39:14.512 ut , 257.948 s after the bat trigger ( taking into account the 3.564 s propagation delay from _ swift _ to _ wind _ ) .
it was detected by the s2 detector which observes the north ecliptic hemisphere ; the incident angle was 63@xmath352 .
count rates are continuously recorded by _ konus - wind _ in three energy bands : g1(1870 kev ) , g2(70300 kev ) , and g3(3001150 kev ) .
data collected in this `` waiting mode '' are acquired in 2.944 s timing bins . the time history recorded in the three energy ranges
can be considered a continuous three - channel spectrum . immediately following the _ konus - wind _ trigger , the instrument began simultaneously collecting data in `` trigger '' mode , as well . from @xmath34 to @xmath36 s , 64 spectra , each composed of 101 energy channels ranging from 18 kev
14 mev , were accumulated .
the time resolution of these `` trigger mode '' spectra varies from 64 ms
8.192 s and is determined by an automated on - board algorithm based on count rate .
data were then processed using standard _ konus - wind _ analysis tools and spectra were fit with xspec .
at least five pulses are evident in the _ konus - wind _ light curve ( labeled @xmath37 , see fig .
[ fig : early_all]b and table [ tab : kwpeaks ] ) .
peak @xmath25 , which generated the bat trigger , appears to be a weak precursor .
figure [ fig : kwhr ] shows the three - channel light curve of grb050820a , as well as the hardness ratios . with the exception of the precursor ( peak @xmath25 ) , the burst shows an overall hard - to - soft evolution over the entire burst duration , as well as within some of the individual peaks ( @xmath26 and @xmath38 ) .
the spectra of individual pulses are well fit by a cut - off power - law model : @xmath39 ; here @xmath40 is the photon index and @xmath41 is the peak energy of the @xmath42 spectrum ( see e.g. [ sec : early ] ) .
fitting the overall _ konus - wind _
`` trigger '' mode spectrum , accumulated from @xmath43 ( @xmath44 ) s ; this tail contains less than 5% of the total burst fluence . ] in the 182000 kev range yields @xmath45 , and a peak energy @xmath46 kev ( 90% c.l . ; @xmath47 for 62 d.o.f . ) .
however , in examining the full _ konus - wind _ light curve ( fig .
[ fig : early_all]b ) , it is clear the above time interval does not include a sizable fraction of the @xmath0-ray emission . to derive the spectral parameters of the time - integrated spectrum over the main part of the grb ( peaks @xmath26@xmath48 ) , we simultaneously fit the three - channel _ konus - wind _ spectrum accumulated from @xmath49 and the overall multichannel spectrum .
we find @xmath50 and a peak energy @xmath51 kev ( 90% c.l . ; @xmath52 for 64 d.o.f . ) .
not surprisingly the peak energy increased significantly , as the beginning of peak @xmath26 was the hardest portion of the entire burst .
we consider this fit the most accurate estimate of the spectral properties of grb050820a .
the fluence and peak flux for each individual episode are shown in table [ tab : kwpeaks ] .
the total fluence received from grb050820a from 201000 kev ( observer frame ) was @xmath53 ergs @xmath54 ( 90% confidence limit ) .
cccc[!b ] @xmath25 & @xmath55 @xmath56 & @xmath57 & @xmath58 + @xmath26 & @xmath59 @xmath60 & @xmath61 & @xmath62 + @xmath38 & @xmath63 @xmath64 & @xmath65 & @xmath66 + @xmath67 & @xmath68 @xmath69 & @xmath70 & @xmath71 + @xmath48 & @xmath72 @xmath73 & @xmath74 & @xmath75 + total & @xmath55 @xmath73 & @xmath76 & + [ tab : kwpeaks ] [ sec : xrt ] the _ swift _ xrt began observations of grb050820a 88 s after the burst trigger . a bright , fading x - ray source was detected by the automated on - board processing routine at @xmath77 , @xmath78 ( 7 error radius ) and reported in real - time @xcite .
data was taken in window timing mode until @xmath79 hr after the burst , when , due to the decreased count rate , photon counting mode was automatically initiated . here
we have independently processed and reduced the xrt data from grb050820a ( for a previous analysis of the early - time x - ray data , see @xcite ) . to reduce the x - ray data , we used the software tools available from the _ swift _ data center .
we followed standard reduction steps , except for taking measures to mitigate the effects of pulse pileup in our spectral analysis . during the steep rise in the x - ray light curve at @xmath80 s , the xrt countrate exceeds the 200 counts s@xmath15 at which pileup can affect the detector response @xcite .
for this segment , we removed the central two pixels from the point source image . at the beginning of photon counting mode , we removed the central four pixels of the photon counting mode image for our spectral analysis , again in order to mitigate the effect of pulse pileup @xcite .
we fit the x - ray spectrum for phases 1a , 1b , and 2 separately ( see [ sec : anal ] for a full discussion of the division of the x - ray light curve into phases ) .
we used an absorbed powerlaw model , and fit both with the column fixed to the galactic value ( @xmath81 @xmath54 ; @xcite ) , as well as with the column floating . for phase 1a ,
the best fit column is consistent with @xmath82 , and there is little difference between fits with the column floating and fixed . for both phases 1b and 2 we find acceptable fits with fixed column , however the best fit columns are @xmath83 @xmath54 and @xmath84 @xmath54 respectively .
therefore we find marginal evidence for absorption in excess of the galactic value ( c.f .
@xcite ) .
the results of our analysis of the x - ray light curve of grb050820a are shown in table [ tab : xrt ] .
for the discussion below we adopt the spectral fits in table [ tab : xrayab ] .
these were used to scale the binned count rates to fluxes and flux density at a nominal energy of 5 kev .
lrrcccc 1a & 0 & 217 & 2.2 @xmath85 0.3 & 1.18 / 11 & @xmath86 & 0.63 / 349 + 1b & 217 & 257 & @xmath87 & @xmath87 & @xmath88 & 1.02 / 749 + 2 & @xmath89 & @xmath90 & @xmath91 & 1.14 / 29 & @xmath92 & 1.00 / 770 + 3 & @xmath90 & @xmath93 & @xmath94 & @xmath87 & @xmath87 & @xmath87 + [ tab : xrayab ] the automated p60 responded to grb050820a and began a pre - programmed sequence of observations starting 3.43 minutes after the _ swift_trigger .
the system is equipped with an optical ccd with a pixel scale of 0.378 / pixel .
images were taken in the kron @xmath95 and @xmath96 and gunn @xmath97 and @xmath98 filters .
all p60 images are processed with standard iraf @xcite routines by an automated reduction pipeline in real - time .
manual inspection of the first images revealed a bright variable source inside the xrt error circle at @xmath99 , @xmath100 ( see figure [ fig : finder ] ) .
this position was promptly reported as the afterglow of grb050820a @xcite , allowing others to obtain high - resolution spectroscopy of the afterglow @xcite .
we continued to monitor the afterglow of grb050820a with p60 for the following 7 nights , until it was too faint for quantitative photometry .
optical photometry of the afterglow was complicated by the presence of two nearby objects : one @xmath101 mag star located @xmath102south - west of the afterglow , and a fainter @xmath103 mag object located only @xmath104 north - east of the afterglow ( see figure [ fig : finder ] , right panel ) . on some nights of poor seeing ,
the full - width at half - maximum of our point - spread function ( psf ) was larger than @xmath105 .
we have therefore performed psf - matched image subtraction using the cpm technique of @xcite on our optical data .
errors were estimated by placing 5 artificial stars with flux equivalent to the afterglow in locations with similar background contamination .
in addition , we have also used aperture photometry ( daophot ) to extract the afterglow flux . on the first night , the afterglow was bright enough to be well - detected either in single images or short co - additions ( @xmath106 s ) . for these images ,
both nearby sources were below our detection limit .
results from aperture photometry and image subtraction were therefore consistent .
we quote our aperture photometry results for these data , as image defects from imperfect psf - subtraction seemed to artificially inflate these errors .
however , on subsequent nights the afterglow flux was either near or below the level of these nearby objects , and we therefore report results from our image subtraction technique .
photometric calibration was performed relative to 20 field stars provided in @xcite .
kron @xmath95 is similar to cousins @xmath95 ( @xmath107 ) and was treated as identical for photometric calibration of these images .
magnitudes from the standard johnson / cousins system were transformed to gunn @xmath97 using the empirical relation from @xcite .
we found , however , a better correlation between our @xmath108-band filter and the cousins @xmath96 ( @xmath109 ) filter than from the transformation to gunn @xmath108 provided in @xcite .
we therefore use @xmath109 in the remainder of this work . for the gunn @xmath98 filter
, we used the two micron all sky survey ( 2mass ; @xcite ) catalog and the optical photometry provided by @xcite to interpolate the spectral energy distributions of 10 sources to the sloan @xmath110 bandpass .
typical rms variations in the calibration sources were @xmath111 mag in @xmath107 and @xmath109 , @xmath112 mag in @xmath97 , and @xmath113 in @xmath110 .
the results of our p60 observations are shown in table [ tab : optical ] .
while the expected galactic extinction in this direction is small ( @xmath114 mag ; @xcite ) , we have incorporated it into our results because of the large wavelength range spanned by our observations ( @xmath115 mag ) .
errors quoted are 1@xmath116 photometric and instrumental errors summed in quadrature . for this and all other optical data in this work , magnitudes are converted to flux densities using the zero - points reported in @xcite .
the _ swift _ uvot automatically slewed to the bat location and began observations only 80 s after the trigger . however , the uvot also becomes inoperable in the saa and therefore does not cover the period of the main @xmath0-ray emission .
the _ swift _
team reduced the early - time uvot data and reported detections in the @xmath117 , @xmath26 , @xmath118 , and @xmath119 bands @xcite . here
we have independently reduced the @xmath118- , @xmath26- , and @xmath117-band uvot data following the recipe outlined in @xcite ( see 3.6 ) . as a check ,
we re - calculated the @xmath26 and @xmath117 zero - points with respect to the field stars from @xcite . while our zero - points are consistent with the ones quoted in @xcite , we found a much larger scatter ( @xmath120 mag vs. @xmath121 mag ) that could not be attributable solely to spread in the field stars .
we have therefore incorporated the resulting zero - point errors for these data points ( as well as a similar value for @xmath118-band ) .
the results of these observations are shown in table [ tab : optical ] .
we triggered target - of - opportunity observations on the 9.2-m _ hobby - eberly telescope _
beginning on the night of 26 august ut .
observations were taken in both @xmath107 and @xmath109 filters .
a second reference epoch was taken on 29 august ut .
image subtraction was performed on the two epochs to remove contamination from nearby sources ( as described in
[ sec : p60 ] ) .
photometric calibration was performed relative to 10 reference objects from @xcite , and the absolute calibration was of similar quality to the p60 dataset .
our results are reported in table [ tab : optical ] . to better constrain the jet - break time , as well as to investigate the properties of the host galaxy of grb050820a
, we triggered our cycle 14 _ hubble space telescope _
program ( go-10551 ; pi : kulkarni ) . using the wide - field camera ( wfc ) channel of the advanced camera for surveys ( acs ) , we imaged the field of grb050820a on 29 september 2005 ut ( @xmath122 d ) in the @xmath123 ( @xmath124 ) , @xmath125 ( @xmath126 ) , and @xmath127 ( @xmath110 ) filters ( see figure [ fig : finder ] ) . a second epoch to study the host galaxy is scheduled for july 2006 , and these results will be reported in a future work . the hst data were processed using the ` multidrizzle ` routine @xcite in the ` stsdas ` iraf package .
we used ` pixfrac`@xmath128 and ` pixscale`@xmath129 for the drizzling procedure , resulting in a pixel scale of @xmath130 / pixel .
the astrometry on these images was then tied to a p60 co - addition of all the @xmath107-band data taken on 20 august 2005 ( which is itself tied to the usno - b1 astrometric grid ) .
the afterglow is well separated from any nearby objects in the field , and so we have followed the recipe for aperture photometry from @xcite . as a note of caution , however , flux from an underlying host galaxy could affect the results reported here .
final values for the late time afterglow flux will require image subtraction of any host galaxy contribution , expected following our july observations . @xmath123 and @xmath125 magnitudes were converted to the @xmath107 and @xmath109 band - passes using synthetic spectra from table 22 of @xcite .
the results of our measurements are shown in table [ tab : optical ] .
[ sec : vla ] in table [ tab : radio ] we summarize our radio observations of grb050820a , spanning 0.161 days after the explosion .
all observations were conducted with the very large array in standard continuum mode with a bandwidth of 2 @xmath131 50mhz centered at 4.86 , 8.46 or 22.5ghz .
the array was in the @xmath38 configuration , with an angular resolution of 2.3 .
we used 3c48 ( j0137@xmath132331 ) for flux calibration , while j2212 + 239 and j2225 + 213 were used to monitor phase .
data were reduced using standard packages within the astronomical image processing system ( aips ) .
rrrc aug 20.39 & 0.116 & 4.86 & @xmath133 + aug 20.39 & 0.116 & 8.46 & @xmath134 + aug 20.39 & 0.116 & 22.5 & @xmath135 + aug 21.20 & 0.93 & 8.46 & @xmath136 + aug 22.42 & 2.15 & 4.86 & @xmath137 + aug 22.42 & 2.15 & 8.46 & @xmath138 + aug 22.42 & 2.15 & 22.5 & @xmath139 + aug 24.38 & 4.11 & 4.86 & @xmath140 + aug 24.38 & 4.11 & 8.46 & @xmath141 + aug 25.32 & 5.05 & 8.46 & @xmath142 + aug 26.40 & 6.13 & 8.46 & @xmath143 + aug 28.37 & 8.10 & 8.46 & @xmath144 + sep 1.33 & 12.06 & 8.46 & @xmath145 + sep 4.18 & 14.91 & 8.46 & @xmath146 + sep 15.20 & 25.93 & 8.46 & @xmath147 + oct 20.19 & 60.92 & 8.46 & @xmath148 + [ tab : radio ]
in this section we provide an analysis of the x - ray , optical , and radio light curves and spectra of the afterglow of grb050820a . we have divided the x - ray and optical light curves into segments ( phases 14 ) based on noticeable temporal breaks .
we then investigated each segment independently , fitting the light curve and spectra to power - law decay indices of the form : @xmath149 .
the lack of a bright afterglow makes such analysis impossible in the radio .
the x - ray and optical light curves , with temporal divisions marked as dashed vertical lines , are shown in figure [ fig : xolc ] .
phase 1 begins with the bat trigger and ends with the resumption of x - ray observations at @xmath150 s. the x - ray and optical data behave differently in phase 1 , resulting in unique subdivisions for the two band - passes .
however , phase 1 is the only epoch to show such divergent behavior . with the emergence of _ swift _ from the saa at @xmath151 , both the x - ray and optical light curves
exhibit a relatively shallow decline .
this characterizes phase 2 , which ends when the x - ray decline steepens at @xmath152 s. the decay slope in phase 3 is steeper than in phase 2 in both band - passes .
phase 3 extends out to the last x - ray detection at @xmath153 s. between this time and the hst optical observations , the optical decay must have significantly steepened .
this last epoch , with only optical data , we define as phase 4 . the results of this power - law analysis are shown in tables [ tab : xrayab ] and [ tab : optab ] .
each bandpass is discussed in further detail below .
lrrccc 1a & @xmath154 & 626 & @xmath155 & & 3.00 / 13 + 1b & 626 & @xmath89 & @xmath156 & @xmath157 & 1.53 / 14 + 2 & @xmath89 & @xmath90 & @xmath158 & @xmath159 & 5.7 / 31 + 3 & @xmath90 & @xmath160 & @xmath161 & @xmath87 & 1.43 / 18 + 4 & @xmath93 & @xmath162 & @xmath163 & @xmath87 & @xmath87 + [ tab : optab ] the x - ray light curve of grb050820a is shown in figure [ fig : xolc ] ( top panel ) . in phase 1 , we see two distinct behaviors : initially the x - ray light curve falls rapidly with a decay slope @xmath164 .
this continues until @xmath165 s , and we define this as phase 1a@xmath166 .
the x - ray emission then rises rapidly ( @xmath167 s ; phase 1b@xmath166 ) , after which _ swift
_ enters the saa .
it is clear from the correlation between the @xmath0-ray and x - ray emission in this epoch that the two are related ( see fig . [
fig : early_all]c ) .
a full study of the properties of phase 1b@xmath166 is left to [ sec : early ] . following _ swift _ s emergence from the saa , the light curve in phase 2 shows evidence for a shallower epoch of decline . with the large gap in coverage
, we can not constrain when this transition occurs .
we therefore define @xmath151 to coincide with the resumption of xrt observations at @xmath168 s. a similar break in the optical light curve is also seen near this time ( [ sec : optlc ] ) .
the x - ray data after @xmath151 are not well fit by a single power - law decay ( @xmath169 ; 31 d.o.f . ) , due mostly to a steepening of the decay at @xmath170 s. fitting a broken power - law model to this data , we find an acceptable fit with @xmath171 s ( @xmath172 ; 25 d.o.f . ) .
the resulting decay index before the break ( phase 2 ) is @xmath173 . for phase 3 , we find @xmath174 .
the x - ray spectral index in phase 1a@xmath166 is relatively steep : @xmath175 .
the spectrum hardens significantly in phase 1b@xmath166 ( @xmath176 ) , further justifying our decision to split phase 1 into two separate x - ray segments . in phase 2 ,
the spectrum softens again , to @xmath177 .
there are too few x - ray counts in phase 3 to meaningfully constrain the spectrum .
the results of our analysis of the x - ray data set are summarized in table [ tab : xrayab ] .
the optical light curve from grb050820a is shown in figure [ fig : xolc ] ( bottom panel ) .
phases 1 ( @xmath178 ) , 2 ( @xmath179 ) , and 3 ( @xmath180 ) of the optical light curve have already been defined in terms of the x - ray decay . however , unlike the x - ray , the earliest optical observations indicate the afterglow was getting brighter with time @xcite .
this rise continues until a peak at @xmath181 s , marking the end of phase 1a@xmath182 .
after the peak , the optical light curve in all four p60 filters decays steadily with @xmath183 until @xmath184 .
we note that for this and all subsequent phases , we have constrained @xmath40 to be identical in all optical filters .
a more thorough discussion of phase 1 is left to [ sec : early ] .
much like the x - ray , the optical decay in phases 2 and 3 is poorly fit by a single power - law ( @xmath185 ; 49 d.o.f . ) . in phase 2 ,
the optical decay noticeably flattens ( @xmath186 )
. a much higher degree of variability is seen in the different filters , resulting in a poor fit statistic .
after @xmath187 the decay in phase 3 is again steeper and more uniform , with @xmath188 .
it is clear that if we extrapolate the decay from phase 3 out to the hst observations , the late - time flux is greatly overestimated .
we conclude therefore a break has occurred in the light curve sometime after @xmath189 , and thus we define phase 4 to span @xmath190 .
we estimate the temporal decay in phase 4 to be @xmath191 . due to the limited spectral coverage of our observations , we are unable to provide meaningful constraints on the spectral index @xmath192 in phases 1a@xmath182 , 3 , and 4 . for the remaining epochs
, we have excluded the @xmath118- and @xmath26-band data from our spectral fits , as these are expected to lie below the ly-@xmath40 absorption edge at this redshift .
we attempted to solve for the host galaxy reddening ( @xmath193(host ) ) using extinction laws for the milky way , large and small magellanic clouds from @xcite . in all cases ,
we find a host extinction consistent with zero .
while this result is inconsistent with the column density derived from high - resolution spectroscopy @xcite , it is in agreement with the low host extinction values seen in almost all well - sampled pre-_swift _
afterglows @xcite .
ignoring host reddening , we find @xmath194 and @xmath195 .
while the optical spectrum appears to have steepened in phase 3 , the poor fit quality of this phase precludes any firm conclusions from being drawn .
the results of our analysis of the optical data set are shown in table [ tab : optab ] .
the radio emission rises to a peak sometime around 1 d after the burst ( see figure [ fig : radio ] and table [ tab : radio ] ) .
the radio spectrum at early times is quite chaotic , transitioning through a peak around 8.5 ghz at @xmath196 d to optically thin 4.11 d after the burst .
we note that some of the variation at early times may be due to inter - stellar scintillation @xcite as has been in seen in many other radio afterglows ( see e.g. @xcite ) . the late - time ( @xmath197 d ) radio data show no sign of any afterglow brighter than 200 @xmath198jy .
this is in marked contrast to the bright optical and x - ray afterglows , and is one of the most difficult aspects of the afterglow to account for ( [ sec : cbm ] ) .
in this section , we use the results from our previous analysis to try and explain the broadband emission from grb050820a in the context of the standard fireball model ( see e.g. @xcite for a review ) . in this model ,
highly - relativistic ejecta are emitted by a central engine as shells with varying lorentz factors .
shock fronts formed between these shells produce the high - energy prompt emission ; these collisions are known as internal shocks .
emission from internal shocks is highly non - thermal and results in an ( empirically determined ) power - law spectrum with a cut - off exponential : @xmath199 @xcite .
as the relativistic shells expand and slow down , they eventually encounter the circum - burst medium . again a collision - less shock front forms , accelerating electrons to a power - law distribution of lorentz factors with exponent @xmath200 and minimum lorentz factor @xmath201 .
it is assumed that a constant fraction of the total energy density is partitioned to the electrons ( @xmath202 ) and the magnetic field ( @xmath203 ) .
these accelerated electrons then emit synchrotron radiation , powering the long - lived x - ray , optical , and radio afterglow ( the forward , external shock ) .
additional emission can be generated from shock heating of the ejecta ( the reverse shock ) , leading to a rapidly - decaying flare in the optical and radio light curve @xcite .
the observed afterglow spectrum depends on the relative ordering of the three critical frequencies : @xmath204 , the frequency where self - absorption becomes important , @xmath205 , the characteristic frequency of the emission , and @xmath206 , the frequency above which electrons are able to cool efficiently through radiation .
typical afterglow observations occur when @xmath207 ( i.e. slow cooling ) , resulting in the following spectral indices @xcite : @xmath208 the light curve produced by such emission depends on the radial profile of the circum - burst medium into which the shock is expanding .
the simplest circum - burst medium to consider is one in which the density is constant ( @xmath209 ) .
this scenario is also referred to as an inter - stellar medium ( ism ) . in this case , the flux density will scale as @xcite : @xmath210 alternatively , if we eliminate @xmath200 from the above equations , we find a characteristic relation between @xmath40 and @xmath192 in each spectral regime known as a `` closure relation '' @xcite : @xmath211 the long - soft class of grbs , however , is thought to arise from the deaths of massive stars , as they collapse to form black - holes : the so - called `` collapsar '' model @xcite . in the late stages of evolution ,
massive wolf - rayet stars are stripped of their outer envelopes in a wind , leaving behind a signature @xmath212 density profile that should be discernible in the afterglow light curve .
the analogous temporal decay indices for a wind - like medium are @xcite : @xmath213 the derived closure relations are : @xmath214 the above temporal decay indices and closure relations ( equations [ eqn : t1a]-[eqn : c2d ] ) are only valid for spherically symmetric emission .
grbs , however , are thought to be beamed events . at early times , observers only notice emission from a narrow cone ( opening angle @xmath215 , where @xmath216 is the lorentz factor of the expanding shock ) due to relativistic beaming ( see e.g. @xcite ) .
as the shock slows , however , lateral spreading of the jet becomes important , and the observer eventually notices `` missing '' emission from wider angles @xcite .
this hydrodynamic transition manifests itself as an achromatic steepening in the afterglow light curve , with an expected post - jet - break decay proportional to @xmath217 . with the above formulation in hand ,
we now wish to understand the physical implications of our previous analysis .
the most striking feature of the @xmath0-ray light curve of grb050820a is the large gap between the initial pulse that triggered the _
swift _ bat ( peak a in fig .
[ fig : early_all]b ) and the bulk of the high - energy emission ( @xmath218 s ) .
the natural question arises as to whether this `` precursor '' in grb050820a is then clearly a precursor . ]
results from the same physical mechanism as the bulk of the high energy emission .
many models predict a high - energy component distinct from the prompt grb at early times .
possible mechanisms include the transition from an optically thick to an optically thin environment in the fireball itself @xcite , or the interaction of a jet with a progenitor , presumably a collapsing wolf - rayet star @xcite .
we can securely rule out both of these models for the precursor of grb050820a , as both predict a thermal spectrum .
fitting a thermal model to the bat precursor spectrum results in a fit statistic of @xmath219 ( 75 d.o.f . ) , while a non - thermal power - law model provides an excellent fit ( @xmath220 for 75 d.o.f . ;
@xmath221 ) .
while the precursor may be non - thermal , it is noticeably softer than the majority of the remaining prompt emission ( see figs .
[ fig : kwhr]d and [ fig : kwhr]e ) . a search for precursors in a sample of long , bright _ batse _
bursts revealed such a soft , non - thermal component in a sizable fraction ( 2025% ) of these events @xcite .
furthermore , two of the longest , brightest _ swift _
bursts observed to date , grb041219 @xcite and grb060124 @xcite show a faint , soft precursor followed by a large time lag ( 570 s in the case of grb060124 ) .
these soft precursors are inconsistent with the main prompt emission in most grbs , which exhibits a hard - to - soft evolution in the @xmath0-ray spectrum @xcite .
the @xmath0-ray light curve of grb050820a conforms to this trend only if we ignore the precursor .
furthermore , it is difficult to conceive of a scenario by which internal shocks can generate such long periods of quiescence in a sustained outflow .
the large time lag , soft nature , and repeated occurrence of these precursors hint that they are in fact due to a different emission mechanism than the internal dissipation thought to power the bulk of the high - energy emission .
however , we can not state this conclusively , as would be the case if the precursors were thermal .
if we assume a different emission mechanism , the prompt emission did not begin until 222 s after the _ swift _ trigger . this seemingly small discrepancy in defining @xmath19 affects the calculated temporal decay indices , particularly during the early afterglow ( [ sec : cbm ] ) .
for all temporal decay indices calculated in this work , we consider @xmath19 to coincide with the beginning of the bulk of the high - energy emission ( i.e. 06:38:35 ut on 20 august 2005 ) .
finally , we consider the early x - ray emission .
the temporal decay slope at early times ( @xmath222 ) is too steep to be attributed to a standard forward shock afterglow .
the most popular explanation for the rapid decline of early x - ray emission in _ swift _ grbs is `` high - latitude emission '' : prompt emission from large angles ( @xmath223 ) that , due to relativistic beaming effects , reaches the observer at late times ( @xmath224 , @xcite ) .
however , this results in a well - defined relationship between the spectral and temporal indices ( @xmath225 ) which is inconsistent with the observed values for grb050820a .
@xcite discuss possible mechanisms that could cause the early - time decay slope to deviate from this behavior .
the most realistic possibility is if the x - rays were below the cooling frequency at this very early epoch .
then the closure relation would take the form @xmath226 @xcite , in good agreement with the observed value .
lrrccccrcc 1 & 80.949 & 85.376 & @xmath227 & @xmath87 & @xmath87 & @xmath87 & @xmath117 & @xmath228 & @xmath229 + 2 & 257.839 & 8.448 & @xmath230 & @xmath231 & @xmath232 & 0.83 / 62 & @xmath107 & @xmath233 & 12.5 + 3 & 297.775 & 59.904 & @xmath234 & @xmath87 & @xmath87 & @xmath87 & @xmath109 & @xmath235 & @xmath236 + 4 & 389.167 & 57.344 & @xmath237 & @xmath238 & @xmath239 & 0.78 / 62 & @xmath110 & @xmath240 & 10.0 + 5 & 487.471 & 57.344 & @xmath241 & @xmath242 & @xmath87 & 1.01 / 58 & @xmath107 & @xmath243 & 9.2 + 6 & 602.159 & 49.152 & @xmath244 & @xmath87 & @xmath87 & @xmath87 & @xmath109 & @xmath245 & @xmath246 + 7 &
684.079 & 57.344 & @xmath247 & @xmath87 & @xmath87 & @xmath87 & @xmath110 & @xmath248 & @xmath249 + [ tab : early ] given the long duration and bright fluence , grb050820a provides a rare opportunity to study contemporaneous emission in the optical , x - ray , and @xmath0-ray band - passes . in figure
[ fig : early_all ] , we show the early time ( @xmath250 s ) emission in x - rays ( c ) and optical ( d ) overlaid onto the high - energy light curve of grb050820a .
a look at the x - ray data in figure [ fig : early_all]c shows a strong correlation between the x - ray and @xmath0-ray light curves .
the x - ray light curve , previously in the midst of a decline , abruptly jumps in sync with the high energy emission at @xmath251 s ( phase 1b@xmath252 ) .
in addition to temporal similarities , the x - ray photon index at this epoch ( @xmath253 ) is much harder than at any other epoch in the x - ray light curve , and similar to that derived from the bat ( @xmath254 ) .
thus motivated , we have performed a joint fit of the bat and xrt spectra at this epoch .
unfortunately _ konus - wind _ had yet to trigger , and so no high - energy multi - channel spectra are available from that instrument .
we find both band - passes well - fit by a single power - law with index @xmath255 ( @xmath256 ; 391 d.o.f . ) .
the resulting un - folded spectrum is shown in figure [ fig : batxrt ] .
we conclude the x - rays in phase 1b@xmath252 are generated by the same mechanism as the prompt emission .
it is clear from figure [ fig : early_all]d that , unlike in the x - ray band , there is no strong correlation between optical and @xmath0-ray flux from grb050820a . radical spectral evolution would be required in the optical to explain both band - passes as arising from the same emission mechanism .
we consider this scenario highly unlikely and conclude that , at the very least , the dominant contribution to the optical emission in phase 1 has a different origin than the prompt emission .
we next consider if our optical observations in phase 1 can be explained solely in terms of the standard afterglow formulation .
we have attempted to fit both a simple broken power - law @xcite as well as an analytic solution for the flux density near the optical peak @xcite to our @xmath107- , @xmath109- , and @xmath110-band early - time data .
the resulting fit quality is quite poor ( @xmath257 ; 14 d.o.f . ) with the dominant contribution coming from the data in phase 1a@xmath182 ( i.e. before the peak ) . in spite of the poor fit
, we have included the results for phase 1a@xmath182 in table [ tab : optab ] for reference .
this result is not unexpected , as @xcite have shown that contemporaneous optical imaging of grb050820a with the _ raptor _
telescope can be well - fit as the sum of two independent components : one representing the forward shock and another proportional to the high - energy prompt emission .
we attempted an analogous fit with the p60 and _ konus - wind _ data set . while a better fit statistic ensues , we still do not find an acceptable result ( @xmath258 ; 14 d.o.f . ) .
we conclude that the relatively sparse time sampling of our observations , coupled with the frequent filter changes , make it impossible to verify this result .
independent of any correlation between the prompt optical and @xmath0-ray emission , we note that the decay after @xmath259 s is dominated by the forward shock .
unlike the bright , early - time emission seen from grb990123 @xcite , we see no evidence for rapidly - decaying ( @xmath260 ; @xcite ) reverse - shock emission from an optical flare . finally , using the combination of optical and @xmath0-ray data , we consider the broadband spectral energy distribution ( sed ) of grb050820a at early times . for each of the contemporaneous optical observations , we have extracted fluxes and ( where possible ) spectra from the corresponding _ konus - wind _ observations ( figure [ fig : earlysed ] and table [ tab : early ] ) . following the method of @xcite , we have calculated the @xmath0-ray - to - optical color index , @xmath261 , or lower limits , for each interval .
the ratio varies significantly over the course of our observations .
the value of @xmath262 in interval 2 is consistent with that seen from grb041219a , while later intervals are even brighter in the optical .
in fact , the optical - to-@xmath0-ray flux ratio in interval 5 is over 240 times larger than that observed for grb050401 @xcite .
evidently a large diversity exists in the broadband seds of grbs at early times .
the majority of xrt light curves observed to date have exhibited a period of shallow decline ( @xmath263 ) that is inconsistent with the standard afterglow formulation @xcite .
two models have been invoked to explain this phase , both of which involve injecting energy into the forward shock at late times ( @xmath8 ; see e.g. @xcite for a review ) . in the first
, the central engine is active for long time periods , @xmath8 .
the late - time emission of highly relativistic material injects additional energy into the forward shock , flattening the decay slope @xcite . alternatively ,
towards the end of the @xmath0-ray emission , the central engine may inject material with a smooth distribution of ( decreasing ) lorentz factors .
slower moving material will catch up with the forward shock when it has swept up enough circum - burst material , resulting in a smooth injection of energy at late times @xcite .
while both models explain the flattening of the xrt light curves , they provide different constraints on the nature of the central engine . in the first ( long - lived central engine ) model , the central engine s luminosity , @xmath264 ,
is characterized as @xmath265 this results in the following spectral and temporal power - law indices for a constant - density medium : @xmath266 for a wind - like medium , the analogous results are @xmath267 the refreshed shock scenario is parameterized in terms of the amount of mass ejected with lorentz factor greater than @xmath0 : @xmath268 for the circum - burst profiles considered here , we can define a new variable , @xmath269 , such that we reproduce identical afterglow behavior to that of equations [ eqn : c3a]-[eqn : c3c ] or equations [ eqn : c4a]-[eqn : c4c ] by simply substituting @xmath269 for @xmath270 .
@xmath269 is related to the mass ejection parameter @xmath271 by the following equations @xcite : @xmath272
while the x - ray decay in phase 2 is not as flat as that seen in other _ swift _
bursts , the temporal and spectral decay indices are nonetheless inconsistent with the standard afterglow model for @xmath273 ( equations [ eqn : c1d ] and [ eqn : c2d ] ) .
furthermore , the optical light curve shows a flattening during phase 2 , and is inconsistent with both the closure relations in either medium for @xmath274 ( equations [ eqn : c1c ] and [ eqn : c2c ] ) .
we conclude we are therefore seeing a milder version of the energy injection phase present in many _ swift
_ xrt afterglows . for the x - ray data in phase 2 , we find an acceptable fit for the energy injection models only if @xmath275 .
this corresponds to values of @xmath276 ( @xmath277 , @xmath278 ) and @xmath279 .
the optical data in phase 2 are best fit with a constant - density medium and @xmath280 : @xmath281 ( @xmath282 ) and @xmath283 . both the x - ray and the optical fall in the spectral regime we would expect , providing further confidence in this interpretation .
a prediction of the energy injection hypothesis is a bright reverse shock at early times most easily visible in the radio @xcite .
a reverse shock nicely explains the rapid decline in flux at 8.5 ghz from 14 d after the burst .
furthermore , the transition from a spectrum peaked around 8.5 ghz at @xmath196 d to an optically thin radio spectrum at @xmath284 d can be understood as the reverse shock peak frequency , @xmath285 , passing through the radio . since @xmath286
, this should occur well before the forward shock peak frequency reaches the radio bands .
distinguishing between the two theories to explain the energy injection is quite difficult , as both models can be identically parameterized .
progress in this area would require a large sample of bursts with detailed contemporaneous x - ray and optical light curves .
if the refreshed shocks are due to continued engine activity , they should be correlated with the bright x - ray flares seen in some xrt afterglows . on the other hand ,
if the flat decay is caused by slow - moving ejecta , this behavior should be more uniform from burst to burst .
such an analysis is beyond the scope of this work .
we now turn our attention to the issue of the circum - burst medium .
as discussed earlier , the radial profile of the burst environment affects the temporal decay below the cooling frequency ( equations [ eqn : t1a]-[eqn : t1c ] , [ eqn : t2a]-[eqn : t2c ] ) .
in particular , the closure relationships ( equations [ eqn : c1a]-[eqn : c1c ] , [ eqn : c2a]-[eqn : c2c ] ) are sufficiently different that we should be able to distinguish between the competing models for a well - sampled event like grb050820a .
first we examine the x - ray data . as discussed previously ( [ sec : inject ] ) , the x - ray observations in phase 2 require invoking mild energy injection to explain the shallower - than - expected decay for @xmath275 .
if we assume the x - ray spectral index does not change from phase 2 to phase 3 , then we find @xmath287 and @xmath288 satisfy the standard afterglow closure relation for @xmath289 ( equations [ eqn : c1d ] and [ eqn : c2d ] ) .
the corresponding values for the electron index are : @xmath290 , @xmath291 .
unlike the x - ray observations , the optical bands typically probe frequencies below the cooling frequency , where the closure relations are different for different circum - burst media ( equations [ eqn : c1c ] and [ eqn : c2c ] ) .
we have shown already in [ sec : inject ] that the optical data in phase 2 are better fit by a constant - density medium .
we find that a constant - density medium is favored in the optical in phase 1b@xmath182 and phase 3 as well .
the only closure relation satisfied in phase 1b@xmath182 is for an ism with @xmath292 ( equation [ eqn : c1c ] ) .
the resulting @xmath200-values are @xmath293 , @xmath294 .
we note that had we equated the bat trigger time , @xmath154 , with the onset of the burst ( @xmath19 ) , the temporal slope in phase 1b@xmath182 would not have been consistent with any closure relation . in phase @xmath295
, we can not meaningfully constrain the optical spectral slope . however , using the x - ray - to - optical spectral slope in this phase , @xmath296 , we conclude the optical data in this segment still fall below the cooling frequency .
based solely on the temporal decline then , we can rule out a wind - like medium in this phase , the corresponding electron index ( @xmath297 ) would result in a divergent total energy .
while this possibility has been addressed with more complicated electron energy distributions ( see e.g. @xcite ) , we consider this possibility unlikely
. taken together , the x - ray and optical data provide a consistent picture of the forward shock expanding into a constant density medium .
the late - time ( @xmath298 week ) radio observations , however , are inconsistent with this interpretation . for
a constant - density medium , the peak flux density , @xmath299 , should remain constant in time .
this would predict , if we have correctly interpreted the optical peak as the forward shock ( [ sec : early ] ) , a similar peak ( @xmath300 mjy ) in the radio at @xmath301 days ( @xmath302 ) .
this is well above the vla detection limit at this epoch , yet we only measure @xmath303 @xmath198jy .
while the energy injection phase will delay the arrival of @xmath205 in the radio ( @xmath304 ) , our radio limits extend out to two months after the burst .
it would be very difficult , if not impossible , to delay the peak this long .
furthermore , during the energy injection phase , the peak flux increases with time ( @xmath305 ) .
thus we would expect to see rising emission earlier relative to the peak , counteracting the delay of the peak radio flux .
one explanation for the lack of a bright , late - time radio afterglow is an early jet break ( @xmath306 day ) , as was invoked for grb990123 @xcite .
however , we find no evidence for a jet break in the optical or x - ray light curves out to at least 17 days after the burst ( see [ sec : energy ] ) .
another possibility , invoked to explain the relatively low late - time radio flux from grb050904 , is a high ambient density @xcite . in the case of grb050904 , it was argued that the large density raised the self - absorption frequency , @xmath204 , above the radio observing bands .
this greatly suppresses the radio flux , for the spectrum in this regime is proportional to @xmath307 ( eqn .
[ eqn : nu1a ] ) .
there is no evidence in the radio data for an optically thick spectrum , although spectral data is sparse at late times .
furthermore , broadband modeling of this event ( [ sec : energy ] ) rules out a high ambient density for typical values of the micro - physical parameters @xmath308 and @xmath309 .
we therefore consider this explanation unlikely . alternatively ,
a natural explanation for the low radio flux at late times is a wind - like medium . in a wind - like medium ,
the forward shock peak flux density declines in time as @xmath310 .
the decreasing peak flux counteracts the rising synchrotron emission , suppressing any late - time radio data .
this is of course inconsistent with our x - ray and optical data , which strongly favor a constant - density medium .
one can imagine a scenario in which the environment near the burst ( the regime sampled predominantly by the x - ray and optical data ) is approximately constant in density , while the outer regions ( sampled by the radio at later times ) have a wind - like profile .
however , without a physical justification for such a density profile , this remains little more than speculation .
the lack of a bright radio afterglow remains a puzzling aspect of grb050820a .
using the high - energy fluence derived from _ konus - wind _ satellite ( [ sec : kw ] ) , we calculate the total isotropic energy release in the prompt emission was @xmath311 ergs ( assuming a redshift of @xmath312 ; @xcite ) .
this makes grb050820a one of the most energetic events ( in terms of @xmath313 ) for which a redshift has been measured @xcite .
however , only a fraction of the explosion energy is converted into prompt emission via internal dissipation .
the rest remains in the kinetic energy of the outflow , powering the forward shock and hence the afterglow .
we can estimate the kinetic energy of the afterglow ( @xmath314 ) by examining the the x - ray emission at @xmath315 hr @xcite . at this point ,
the x - rays should be above the cooling frequency .
the flux density is then independent of ambient density and only weakly dependent on @xmath203 .
a joint fit of the phase 3 optical and x - ray data ( after the energy injection has stopped and the system has returned to adiabatic expansion ) constrains the electron energy index : @xmath316 . if we take typical values for @xmath202 ( @xmath317 ) and @xmath203 ( @xmath318 ) @xcite , we find that @xmath319 . for an accurate accounting of the total energy emitted by this event ,
however , we must determine the degree of collimation of the emission .
we therefore examine all the temporal breaks in the optical and x - ray light curves to determine which one ( if any ) shows an achromatic steepening to the @xmath217 decay expected from a jet @xcite .
the only plausible candidate is the transition from phase 3 to 4 in the optical light curve .
the steepening here is achromatic ( i.e. it is seen in all three _ hubble space telescope _ filters ) and much too large to be explained solely by the cooling frequency passing through the optical bands ( although this may have occured as well ) .
any contribution from an underlying host galaxy would only further steepen the decay in phase 4 . with only one observation ,
it is impossible to constrain the post - break decay index .
instead , we assume the post - break decay has a power - law index @xmath320 ( see above ) .
we then find @xmath321 d. this result is consistent with our x - ray observations , which put a lower limit on the jet break time of @xmath322 d. we note that the jet break time we have inferred for grb050820a is extremely large . in the host galaxy reference frame
, the break occurs at @xmath323 , a factor of three larger than any jet break seen in the pre-_swift _ era @xcite . in this respect , too , grb050820a is a strong outlier . to convert the jet break times to a range of opening angles
, we use the relation @xcite : @xmath324 here @xmath325 is the fraction of the total energy converted to prompt @xmath0-ray emission .
the only remaining unknown in equation [ eqn : thetatj ] is the ambient density , @xmath326 . using the ratio of the x - ray and optical data , as well as the canonical values of @xmath308 and @xmath309 , we find the density of the burst must be low : @xmath327 @xmath328 .
afterglow modeling of the late - time optical and x - ray data ( phase 3 , after any continued energy injection has ceased and the shock expands adiabatically ) using the technique of @xcite confirms this result : @xmath329 . combining the above results , we find the opening angle is constrained to fall between between 6.8@xmath330 , corresponding to a beaming factor @xmath331 . while the opening angle is large for a long - soft burst , there are several comparable events in the pre-_swift
_ sample @xcite .
the total collimation - corrected energy emitted in @xmath0-rays ( @xmath332 ) from grb050820a is therefore @xmath333 ergs .
the corresponding value for the blast - wave energy ( @xmath334 ) is @xmath335 ergs .
@xcite have demonstrated an empirical relation between @xmath332 and the peak energy of the prompt emission spectrum in the grb rest frame ( @xmath336 ) .
grb050820a is more energetic than any of the 37 bursts considered in their sample ( in terms of @xmath332 ) , and therefore proves an interesting test case for the so - called ghirlanda relation .
using our calculated value of @xmath332 , ghirlanda predicts @xmath337 mev .
this is marginally consistent with the actual value of @xmath338 mev derived from the _ konus - wind _ dataset . in figure
[ fig : energy ] , we plot a histogram of @xmath332 and @xmath334 for the @xmath9 15 long - soft cosmological bursts for which both quantities have been derived .
we have not included the most nearby events ( grb980425 , grb031203 , and grb060218 ) in our analysis , as these events released significantly less energy than the typical cosmological grb @xcite .
@xcite have shown that , with the exception of the most nearby events , the sum of @xmath332 and @xmath334 is clustered around @xmath339 ergs .
grb050820a is clearly an over - energetic exception , an order of magnitude more energetic than this sample .
in fact , it would require the direct conversion of @xmath340 m@xmath341 ( with 100% efficiency ) to release this much energy . finally , it is important to consider how robust our estimates of @xmath314 , @xmath342 , and @xmath326 are given that the standard afterglow model fails to explain the broadband behavior of grb050820a .
we note that the opening angle is relatively insensitive to both variables ( eqn .
[ eqn : thetatj ] ) ; factors of order unity will be greatly reduced by the @xmath343 exponent .
it is difficult to conceive of a long - soft grb environment where the ambient density is less than @xmath344 , and high densities would only increase the opening angle and thus the energy release .
grb050820a joins a select sample of events with simultaneous observations in the @xmath0-ray and optical bands , and an even smaller group with contemporaneous x - ray observations as well .
such events have led to fundamental advances in our understanding of grbs , including the discovery of a reverse - shock optical flash from grb990123 @xcite and possibly from grb050904 @xcite .
the early - time optical emission from grb041219a also showed a bright optical flash , but the rise was correlated with an accompanying peak of @xmath0-ray emission , suggesting a common origin for the two components @xcite .
the early behavior of grb050820a is unlike either of these events . @xcite
have shown the contemporaneous optical emission is well described as the sum of two components : one proportional to the prompt @xmath0-ray emission and one smoothly varying forward shock term . while the @xmath0-ray component is important for @xmath345 s , the optical peak at @xmath181 s is dominated by emission from the external shock region . furthermore , the post - peak decay rate is inconsistent with reverse - shock emission .
instead we interpret this as the forward shock peak frequency passing through the optical bands .
this is not unlike what was seen in the optical for grb060124 @xcite , although the time resolution in the prompt phase was much poorer than for this burst .
the contemporaneous optical light curves of grb050319 @xcite and grb050401 @xcite did not show this peak phase , but extrapolations to late - times were consistent with the adiabatic expansion of a forward shock .
another different behavior was seen in the early optical light curve of grb050801 , which showed an extended plateau phase correlated with the x - ray emission , hinting at continued energy injection from a central engine refreshing the external shocks @xcite .
the contemporaneous x - ray emission , on the other hand , is the low - energy tail of the prompt emission .
this behavior was also seen for grb060124 @xcite and has been hinted at in the rapid decline in early x - ray light curves attributed to high - latitude emission @xcite , as well as the bright x - ray flares seen in many xrt light curves @xcite .
it is clear then , that , in marked contrast to the x - ray emission , contemporaneous optical emission exhibits a large diversity in behavior .
unfortunately the physical mechanism behind this diversity remains to be explained .
the issue of burst geometry is a particularly interesting one in the _ swift _ era .
the steep post - break decay slope , seen simultaneously in multiple filters , makes grb050820a one of the most convincing examples of a beamed event in the _ swift _ sample .
the x - ray afterglow , however , is too faint at late times to provide broadband confirmation .
in fact , very few _ swift _ bursts , including those , like grb050408 @xcite , that have been followed for months , show signs of a jet break in the xrt light curve @xcite .
typical jet breaks in pre-_swift _
bursts occurred on time - scales of several days @xcite .
@xcite predicted that _ swift _ would detect bursts with wider opening angles than previous missions due to the increased sensitivity of the bat .
however , not a single _ swift _ afterglow has shown a convincing jet - break - like transition in multiple band - passes ( candidates include grb050525a and grb050801 ; see @xcite and references therein for a more thorough discussion ) . while it may be that most jet breaks , like grb050820a , occur at late times , beyond the sensitivity of the xrt and most ground - based facilities , this is nonetheless troubling . on the one hand , few if any x - ray jet breaks were seen in pre-_swift _
bursts ; all collimation angles were determined from optical ( and sometimes radio ) light curves .
conversely , given the large number of well - sampled xrt light curves , and the fact that such fundamental results for grb cosmology as the ghirlanda relation rest on our picture of grbs as a - spherical events , this is clearly a matter that merits further investigation .
one consequence of the large opening angle associated with grb050820a is a correspondingly large burst and afterglow energy .
in fact , of all the bursts compiled in the @xcite sample , grb050820a has the largest prompt energy release . and unlike grb990123 @xcite
, this large @xmath0-ray energy was accompanied by a correspondingly large kinetic energy imparted to the afterglow .
the only comparable event for which such energies could be determined was the high - redshift grb050904 , which released a total energy of @xmath346 ergs @xcite . given the large @xmath0-ray fluence , similar events should have been easily detected by both _ swift _ and previous grb missions . and given the bright optical afterglow and the late jet - break , such events are strongly favored for ground - based follow - up ( i.e. redshift determination ) .
the lack of a large sample of such events means they must be relatively rare in the universe . like many other _ swift _
grbs , the x - ray light curve of grb050820a exhibits a phase of shallow decay incompatible with the standard forward shock model @xcite .
grb050820a is relatively unique , however , in that this epoch is also well - sampled in the optical .
the seemingly simultaneous breaks in the optical light curve bolster the commonly - held belief that this phase is caused by some form of refreshed shocks @xcite .
coupled with the large gap between the precursor and the bulk of the prompt emission , the late - time energy injection poses fundamental challenges to any central engine model . finally , we return to the question of the radio afterglow .
radio observations typically probe low - lorentz factor ejecta ( @xmath347 ) at large distances from the central engine ( @xmath348 cm ) . the forward shock peak frequency reaches the radio much later than the optical or x - ray bands .
thus , radio emission is usually visible at later times than optical or x - ray emission , and is well - suited to study afterglows when the emission is isotropic ( i.e. after the jet break ) or even in some cases when the ejecta has slowed to newtonian expansion @xcite . for this reason
, late - time radio observations are considered the most accurate method for model - independent calorimetry .
for grb050820a , this paradigm has broken down .
the burst had a bright optical and x - ray afterglow , but weak emission in the radio .
it is hoped that further studies of such energetic grbs in the _ swift _ era will help to elucidate some of these puzzles . ,
d. 1986 , in instrumentation in astronomy vi ; proceedings of the meeting , tucson , az , mar .
4 - 8 , 1986 .
part 2 ( a87 - 36376 15 - 35 ) .
bellingham , wa , society of photo - optical instrumentation engineers , 1986 , p. 733 . ,
d. l. crawford , 733+ lrrrc aug 20 06:36:25 & 92.0 & 10.0 & 0.90 & @xmath349 + aug 20 06:36:35 & 102.0 & 10.0 & @xmath87 & @xmath350 + aug 20 06:36:45 & 112.0 & 10.0 & @xmath87 & @xmath351 + aug 20 06:36:55 & 122.0 & 10.0 & @xmath87 & @xmath352 + aug 20 06:37:05 & 132.0 & 10.0 & @xmath87 & @xmath353 + aug 20 06:37:15 & 142.0 & 10.0 & @xmath87 & @xmath354 + aug 20 06:37:25 & 152.0 & 10.0 & @xmath87 & @xmath355 + aug 20 06:37:35 & 162.0 & 10.0 & @xmath87 & @xmath356 + aug 20 06:37:45 & 172.0 & 10.0 & @xmath87 & @xmath357 + aug 20 06:37:55 & 182.0 & 10.0 & @xmath87 & @xmath358 + aug 20 06:38:05 & 192.0 & 10.0 & @xmath87 & @xmath359 + aug 20 06:38:15 & 202.0 & 10.0 & @xmath87 & @xmath360 + aug 20 06:38:25 & 212.0 & 10.0 & @xmath87 & @xmath361 + aug 20 06:38:35 & 222.0 & 10.0 & -0.10 & @xmath362 + aug 20 06:38:45 & 232.0 & 10.0 & @xmath87 & @xmath363 + aug 20 06:38:55 & 242.0 & 10.0 & @xmath87 & @xmath364 + aug 20 06:39:05 & 252.0 & 10.0 & @xmath87 & @xmath365 + aug 20 07:56:43 & @xmath366 & 250.0 & 1.20 & @xmath367 + aug 20 08:00:53 & @xmath368 & 250.0 & @xmath87 & @xmath369 + aug 20 08:05:03 & @xmath370 & 250.0 & @xmath87 & @xmath371 + aug 20 08:09:13 & @xmath372 & 250.0 & @xmath87 & @xmath373 + aug 20 08:13:23 & @xmath374 & 250.0 & @xmath87 & @xmath375 + aug 20 11:41:43 & @xmath376 & @xmath377 & @xmath87 & @xmath378 + aug 20 12:23:23 & @xmath379 & @xmath377 & @xmath87 & @xmath380 + aug 20 13:05:03 & @xmath381 & @xmath377 & @xmath87 & @xmath382 + aug 20 14:28:23 & @xmath383 & @xmath377 & @xmath87 & @xmath384 + aug 20 15:10:03 & @xmath385 & @xmath377 & @xmath87 & @xmath386 + aug 20 15:51:43 & @xmath387 & @xmath377 & @xmath87 & @xmath388 + aug 20 16:33:23 & @xmath389 & @xmath377 & @xmath87 & @xmath390 + aug 20 17:15:03 & @xmath391 & @xmath377 & @xmath87 & @xmath392 + aug 20 17:56:43 & @xmath393 & @xmath377 & @xmath87 & @xmath394 + aug 20 19:20:03 & @xmath395 & @xmath377 & @xmath87 & @xmath396 + aug 21 15:48:01 & @xmath397 & @xmath398 & @xmath87 & @xmath399 + aug 21 18:34:41 & @xmath400 & @xmath398 & @xmath87 & @xmath401 + aug 21 21:21:21 & @xmath402 & @xmath398 & @xmath87 & @xmath403 + aug 22 00:08:01 & @xmath404 & @xmath398 & @xmath87 & @xmath405 + aug 22 02:54:41 & @xmath406 & @xmath398 & @xmath87 & @xmath407 + aug 23 14:30:22 & @xmath408 & @xmath398 & @xmath87 & @xmath409 + aug 23 17:17:02 & @xmath410 & @xmath398 & @xmath87 & @xmath411 + aug 24 18:02:47 & @xmath412 & @xmath413 & @xmath87 & @xmath414 + aug 25 03:46:07 & @xmath415 & @xmath413 & @xmath87 & @xmath416 + aug 27 19:38:46 & @xmath417 & @xmath418 & @xmath87 & @xmath419 + aug 29 06:22:06 & @xmath420 & @xmath413 & @xmath87 & @xmath421 + sep 04 15:45:23 & @xmath422 & @xmath423 & @xmath87 & @xmath424 + sep 05 19:32:03 & @xmath425 & @xmath423 & @xmath87 & @xmath426 + sep 06 23:18:43 & @xmath427 & @xmath423 & @xmath87 & @xmath428 + [ tab : xrt ] lrlrrcr aug 20 08:14:47 & @xmath429 & uvot & @xmath118 & 693.3 & 18.12 @xmath85 0.13 & * + aug 20 12:49:02 & @xmath430 & uvot & @xmath118 & 899.8 & 19.11 @xmath85 0.14 & * + aug 20 16:01:58 & @xmath431 & uvot & @xmath118 & 899.8 & 19.40 @xmath85 0.14 & * + aug 20 19:14:54 & @xmath432 & uvot & @xmath118 & 899.8 & 19.66 @xmath85 0.15 & * + aug 20 22:59:51 & @xmath433 & uvot & @xmath118 & 392.3 & 19.70 @xmath85 0.18 & * + aug 20 09:36:06 & @xmath434 & uvot & @xmath26 & 899.8 & 18.79 @xmath85 0.16 & * + aug 20 13:04:10 & @xmath435 & uvot & @xmath26 & 899.8 & 19.24 @xmath85 0.15 & * + aug 20 16:17:05 & @xmath436 & uvot & @xmath26 & 897.3 & 19.74 @xmath85 0.16 & * + aug 20 19:30:02 & @xmath437 & uvot & @xmath26 & 899.8 & 19.87 @xmath85 0.16 & * + aug 21 23:30:41 & @xmath438 & rtt150 & @xmath26 & 1800.0 & 21.28 @xmath85 0.06 & 1 + aug 22 22:58:17 & @xmath439 & rtt150 & @xmath26 & 7860.0 & 22.05 @xmath85 0.06 & 2 + aug 23 22:09:05 & @xmath440 & rtt150 & @xmath26 & 5400.0 & 22.38 @xmath85 0.08 & 2 + aug 20 07:01:53 & @xmath441 & p60 & @xmath97 & 120.0 & 16.27 @xmath85 0.04 & * + aug 20 07:12:35 & @xmath442 & p60 & @xmath97 & 120.0 & 16.65 @xmath85 0.04 & * + aug 20 07:23:24 & @xmath443 & p60 & @xmath97 & 120.0 & 16.92 @xmath85 0.05 & * + aug 20 08:04:47 & @xmath444 & p60 & @xmath97 & 360.0 & 17.59 @xmath85 0.05 & * + aug 20 08:29:42 & @xmath445 & p60 & @xmath97 & 360.0 & 17.87 @xmath85 0.06 & * + aug 20 08:54:50 & @xmath446 & p60 & @xmath97 & 360.0 & 18.11 @xmath85 0.06 & * + aug 20 09:33:18 & @xmath447 & p60 & @xmath97 & 720.0 & 18.29 @xmath85 0.06 & * + aug 20 10:43:25 & @xmath448 & p60 & @xmath97 & 720.0 & 18.36 @xmath85 0.06 & * + aug 20 11:39:32 & @xmath449 & p60 & @xmath97 & 720.0 & 18.59 @xmath85 0.09 & * + aug 20 06:36:58 & 125.0 & uvot & @xmath117 & 89.0 & 18.18 @xmath85 0.20 & * + aug 20 07:52:58 & @xmath450 & uvot & @xmath117 & 99.8 & 17.20 @xmath85 0.13 & * + aug 20 10:04:28 & @xmath451 & uvot & @xmath117 & 337.0 & 18.16 @xmath85 0.12 & * + aug 20 11:12:34 & @xmath452 & uvot & @xmath117 & 899.8 & 18.39 @xmath85 0.09 & * + aug 20 14:25:30 & @xmath453 & uvot & @xmath117 & 899.8 & 18.86 @xmath85 0.11 & * + aug 20 17:38:26 & @xmath454 & uvot & @xmath117 & 899.8 & 19.35 @xmath85 0.12 & * + aug 20 21:03:50 & @xmath455 & uvot & @xmath117 & 899.8 & 19.14 @xmath85 0.11 & * + aug 21 16:09:34 & @xmath456 & uvot & @xmath117 & 899.8 & 20.24 @xmath85 0.19 & * + aug 21 19:22:30 & @xmath457 & uvot & @xmath117 & 899.8 & 20.40 @xmath85 0.22 & * + aug 20 06:38:49 & 236.0 & p60 & @xmath107 & 60.0 & 15.39 @xmath85 0.04 & * + aug 20 06:43:31 & 517.0 & p60 & @xmath107 & 60.0 & 14.65 @xmath85 0.02 & * + aug 20 06:48:25 & 812.0 & p60 & @xmath107 & 60.0 & 15.06 @xmath85 0.03 & * + aug 20 06:53:59 & @xmath458 & p60 & @xmath107 & 120.0 & 15.42 @xmath85 0.02 & * + aug 20 07:04:34 & @xmath459 & p60 & @xmath107 & 120.0 & 16.05 @xmath85 0.03 & * + aug 20 07:15:19 & @xmath460 & p60 & @xmath107 & 120.0 & 16.39 @xmath85 0.02 &
* + aug 20 07:32:13 & @xmath461 & p60 & @xmath107 & 360.0 & 16.77 @xmath85 0.02 & * + aug 20 08:12:58 & @xmath462 & p60 & @xmath107 & 360.0 & 17.32 @xmath85 0.04 & * + aug 20 08:38:05 & @xmath463 & p60 & @xmath107 & 360.0 & 17.48 @xmath85 0.04 & * + aug 20 08:46:53 & @xmath464 & prompt-5 & @xmath107 & 660.0 & 17.52 @xmath85 0.09 & 3 + aug 20 09:03:19 & @xmath465 & p60 & @xmath107 & 360.0 & 17.69 @xmath85 0.04 & * + aug 20 09:28:50 & @xmath466 & p60 & @xmath107 & 360.0 & 17.85 @xmath85 0.04 & * + aug 20 10:00:44 & @xmath467 & p60 & @xmath107 & 360.0 & 17.78 @xmath85 0.04 & * + aug 20 10:38:59 & @xmath468 & p60 & @xmath107 & 360.0 & 17.97 @xmath85 0.04 & * + aug 20 11:05:54 & @xmath469 & p60 & @xmath107 & 360.0 & 18.01 @xmath85 0.04 & * + aug 20 11:33:14 & @xmath470 & p60 & @xmath107 & 360.0 & 18.03 @xmath85 0.04 & * + aug 20 12:02:31 & @xmath471 & p60 & @xmath107 & 360.0 & 18.16 @xmath85 0.05 & * + aug 21 00:37:53 & @xmath472 & rtt150 & @xmath107 & 900.0 & 19.36 @xmath85 0.01 & 1 + aug 21 04:16:53 & @xmath473 & prompt-5 & @xmath107 & 5370.0 & 19.94 @xmath85 0.31 & 3 + aug 22 00:00:05 & @xmath474 & rtt150 & @xmath107 & 1800.0 & 20.26 @xmath85 0.05 & 1 + aug 22 07:17:03 & @xmath475 & p60 & @xmath107 & 6840.0 & 20.51 @xmath85 0.11 & * + aug 22 23:00:41 & @xmath476 & rtt150 & @xmath107 & 3900.0 & 20.90 @xmath85 0.03 & 2 + aug 23 08:06:11 & @xmath477 & p60 & @xmath107 & 8400.0 & 20.89 @xmath85 0.10 & * + aug 23 22:18:05 & @xmath478 & rtt150 & @xmath107 & 2700.0 & 21.18 @xmath85 0.04 & 2 + aug 24 08:46:32 & @xmath479 & p60 & @xmath107 & 8400.0 & 21.22 @xmath85 0.11 & * + aug 25 09:33:11 & @xmath480 & p60 & @xmath107 & 2880.0 & 21.34 @xmath85 0.13 & * + aug 26 05:05:39 & @xmath481 & het & @xmath107 & 600.0 & 21.57 @xmath85 0.08 & * + aug 26 08:28:20 & @xmath482 & p60 & @xmath107 & 3600.0 & 21.64 @xmath85 0.12 & * + aug 27 08:37:13 & @xmath483 & p60 & @xmath107 & 4800.0 & 21.80 @xmath85 0.12 & * + aug 27 22:49:53 & @xmath484 & rtt150 & @xmath107 & 1500.0 & 22.02 @xmath85 0.10 & 4 + sep 26 01:39:38 & @xmath485 & hst & @xmath123 & 800.0 & 24.59 @xmath85 0.08 & * + & & & @xmath107 & & 24.55 @xmath85 0.08 & * + aug 20 06:40:21 & 328.0 & p60 & @xmath109 & 60.0 & 14.91 @xmath85 0.02 & * + aug 20 06:45:19 & 626.0 & p60 & @xmath109 & 60.0 & 14.42 @xmath85 0.01 & * + aug 20 06:50:06 & 913.0 & p60 & @xmath109 & 60.0 & 14.78 @xmath85 0.01 & * + aug 20 06:56:37 & @xmath486 & p60 & @xmath109 & 120.0 & 15.24 @xmath85 0.02 & * + aug 20 07:07:15 & @xmath487 & p60 & @xmath109 & 120.0 & 15.74 @xmath85 0.02 & * + aug 20 07:18:03 & @xmath488 & p60 & @xmath109 & 120.0 & 16.07 @xmath85 0.02 & * + aug 20 07:40:11 & @xmath489 & p60 & @xmath109 & 360.0 & 16.54 @xmath85 0.03 & * + aug 20 08:21:15 & @xmath490 & p60 & @xmath109 & 360.0 & 17.02 @xmath85 0.05 & * + aug 20 08:46:26 & @xmath491 & p60 & @xmath109 & 360.0 & 17.22 @xmath85 0.04 & * + aug 20 08:46:53 & @xmath464 & prompt-3 & @xmath109 & 1560.0 & 17.31 @xmath85 0.08 & 3 + aug 20 09:11:47 & @xmath492 & p60 & @xmath109 & 360.0 & 17.23 @xmath85 0.05 & * + aug 20 09:37:27 & @xmath493 & p60 & @xmath109 & 360.0 & 17.34 @xmath85 0.03 & * + aug 20 10:21:01 & @xmath494 & p60 & @xmath109 & 360.0 & 17.48 @xmath85 0.04 & * + aug 20 10:47:55 & @xmath495 & p60 & @xmath109 & 360.0 & 17.68 @xmath85 0.04 & * + aug 20 11:14:54 & @xmath496 & p60 & @xmath109 & 360.0 & 17.71 @xmath85 0.05 & * + aug 20 11:42:22 & @xmath497 & p60 & @xmath109 & 360.0 & 17.72 @xmath85 0.04 & * + aug 21 04:04:53 & @xmath498 & prompt-3 & @xmath109 & 5440.0 & 18.33 @xmath85 0.11 & 3 + aug 22 00:07:53 & @xmath404 & rtt150 & @xmath109 & 1800.0 & 19.74 @xmath85 0.07 & 1 + aug 22 07:25:05 & @xmath499 & p60 & @xmath109 & 6840.0 & 19.97 @xmath85 0.12 & * + aug 22 23:06:41 & @xmath500 & rtt150 & @xmath109 & 3900.0 & 20.37 @xmath85 0.05 & 2 + aug 23 08:22:42 & @xmath501 & p60 & @xmath109 & 7560.0 & 20.48 @xmath85 0.12 & * + aug 23 22:24:05 & @xmath502 & rtt150 & @xmath109 & 2700.0 & 20.78 @xmath85 0.09 & 2 + aug 24 09:11:48 & @xmath503 & p60 & @xmath109 & 8400.0 & 20.59 @xmath85 0.11 & * + aug 25 10:44:39 & @xmath504 & p60 & @xmath109 & 2400.0 & 21.02 @xmath85 0.15 & * + aug 26 04:49:33 & @xmath505 & het & @xmath109 & 1200.0 & 21.25 @xmath85 0.10 & * + aug 26 09:27:24 & @xmath506 & p60 & @xmath109 & 4800.0 & 21.17 @xmath85 0.14 & * + aug 27 08:34:02 & @xmath507 & p60 & @xmath109 & 4440.0 & 21.30 @xmath85 0.13 & * + sep 26 02:01:58 & @xmath508 & hst & @xmath125 & 800.0 & 24.32 @xmath85 0.09 & * + & & & @xmath109 & & 24.27 @xmath85 0.09 & * + aug 20 06:41:50 & 417.0 & p60 & @xmath110 & 60.0 & 13.93 @xmath85 0.11 & * + aug 20 06:46:45 & 712.0 & p60 & @xmath110 & 60.0 & 14.27 @xmath85 0.14 & * + aug 20 06:51:50 & @xmath509 & p60 & @xmath110 & 60.0 & 14.62 @xmath85 0.21 & * + aug 20 06:59:16 & @xmath510 & p60 & @xmath110 & 120.0 & 15.02 @xmath85 0.13 & * + aug 20 07:09:56 & @xmath511 & p60 & @xmath110 & 120.0 & 15.49 @xmath85 0.14 & * + aug 20 07:20:45 & @xmath512 & p60 & @xmath110 & 120.0 & 15.96 @xmath85 0.20 & * + sep 26 03:06:14 & @xmath513 & hst & @xmath127 & 1600.0 & 24.09 @xmath85 0.09 & * [ tab : optical ] | we present observations of the unusually bright and long @xmath0-ray burst grb050820a , one of the best - sampled broadband data sets in the _ swift_era .
the @xmath0-ray light curve is marked by a soft precursor pulse some 200 s before the main event ; the lack of any intervening emission suggests that it is due to a physical mechanism distinct from the grb itself .
the large time lag between the precursor and the main emission enabled simultaneous observations in the @xmath0-ray , x - ray , and optical band - passes , something only achieved for a handful of events to date . while the contemporaneous x - rays are the low - energy tail of the prompt emission , the optical does not directly track the @xmath0-ray flux .
instead , the early - time optical data appear mostly consistent with the forward shock synchrotron peak passing through the optical , and are therefore likely the beginning of the afterglow . on hour time scales after the burst , the x - ray and optical light curves are inconsistent with an adiabatic expansion of the shock into the surrounding region , but rather indicate that there is a period of energy injection .
observations at late times allow us to constrain the collimation angle of the relativistic outflow to @xmath1 .
our estimates of both the kinetic energy of the afterglow ( @xmath2 ergs ) and the prompt @xmath0-ray energy release ( @xmath3 ergs ) make grb050820a one of the most energetic events for which such values could be determined . |
in the last years , graphene oxide ( go ) has attracted remarkable interest as a precursor for a large - scale and low - cost production of graphene - based materials and as a fundamental constituent in new functional composite materials for optoelectronics , photovoltaics , and nanobiology devices.@xcite go is obtained by liquid exfoliation of chemically oxidized graphite ( graphite oxide);@xcite at a later stage , a graphene - like material known as reduced graphene oxide ( rgo ) can be derived from go , by removing the oxygen functionalities.@xcite however , after more than 150 years since its discovery , many questions remain open about the exact chemistry and structure of ( r)go at the atomic scale .
@xcite currently , the most acknowledged model for go corresponds to a nonstoichiometric structure with a random arrangement of epoxide ( -o- in bridge position ) and hydroxyl ( -oh ) functionalities at both sides of the carbon basal plane , with small residual nonfunctionalized graphitic areas .
@xcite difficulties in defining the exact relative amount and local arrangement of oxygen groups arise from the strong dependence of go chemistry and structure on the specific graphite precursor and oxidation parameters.@xcite moreover go turned out to be unstable in air and water , resulting in spontaneous reduction and flakes fragmentation.@xcite the uncertainties on go naturally reflect on rgo , whose chemistry and structure also depend on the particular reduction method and parameters.@xcite this lack of knowledge represents a strong limit for the understanding of ( r)go physical properties and hence for a controlled use of these promising functional materials .
new insights could be provided only through the use of improved characterization techniques . up to now
, several spectroscopic techniques have been used to determine the type and amount of oxygen functional groups . in particular ,
the oxygen level has been estimated mainly by elemental analysis , xps and edx on a variety of samples going from graphite oxide powders to ( r)go films and thin flakes .
the oxygen content has been evaluated in a range between about 15 and 35 atomic percent ( at.% ) , i.e. between @xmath15:1 and @xmath12:1 in terms of the c / o atomic ratio . in rgo
the oxygen content can be lowered to a limit of 0.4 at.% ( @xmath1250:1 c / o atomic ratio ) , depending on the particular reduction process.@xcite however , these values can only be considered as spatial average quantifications , that can not account for the irregular chemistry of go . indeed ,
strong structural inhomogeneities at the nanometer scale have been revealed in both go and rgo by high resolution transmission electron microscopy ( hrtem).@xcite decisive evidence on go chemistry could only be derived by spectroscopy at the atomic scale .
electron energy - loss spectroscopy ( eels ) in a scanning transmission electron microscope ( stem ) could in principle be suitable for such analysis , providing elemental quantification by core eels mapping and chemical structure analysis through energy - loss near - edge structures ( elnes ) , down to the atomic scale.@xcite however , the limited use of stem - eels for the study of go and rgo@xcite resides in the strong sensitivity of these materials to high energy electron irradiation .
for instance , atoms mobility and fast amorphization have been clearly demonstrated by time series of hrtem images , at 80 kv acceleration voltage.@xcite illumination sensitivity is indeed highly expected in the case of go and rgo , since single or few layer thick specimens and light atoms are intrinsically more affected by radiation damages , such as knock - on of carbon atoms and radiolysis of oxygen functionalities resulting in mass loss .
irradiation effects in a stem could in principle be strongly reduced by the use of low accelerating voltages , low electron - dose , and sample cooling .
@xcite in this work , thanks to a dedicated experimental setup combining a liquid nitrogen ( ln ) cooling system at the sample stage with a low noise ln cooled ccd camera , we investigate few layer go and rgo by core eels spectrum imaging in a stem microscope in low dose acquisition modes .
we show that the oxygen content in individual ( r)go flakes is strongly heterogeneous at the nanometric scale and that specific near - edge carbon fine structures could be associated with different oxygen contents and bindings , indicating separated chemical phases . finally ,
on the basis of these experimental evidence and complementary dft - based numerical simulations , we suggest a structural model for the highly oxidized regions in go where all carbon atoms are functionalized by -oh groups .
[ [ synthesis . ] ] synthesis .
+ + + + + + + + + + graphene oxide has been synthesized by the modified hummers method,@xcite and successive reduction has been achieved by hydrazine and thermal treatments .
few layer flakes with an average lateral size of a few micrometers have been obtained by dispersion of the dried material in ethanol and ultrasonication .
tem grids were dried in air , first placed in the stem microscope air lock ( @xmath2 mbar ) , and then moved in the microscope column ( @xmath3 mbar ) and cooled down to about 150 k. high angle annular dark field ( haadf ) images show a nonhomogeneous thickness of the flakes , with thinner regions some hundreds of nanometers wide located at the flakes borders .
elemental investigations through core eels indicate a clear dominance of carbon and oxygen with only negligible residues of nitrogen ( in rgo , deriving from the hydrazine reduction agent @xcite ) at few limited regions of the flakes ( about 20 nm wide ) .
[ [ electron - energy - loss - spectroscopy . ] ] electron energy loss spectroscopy .
+ + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + eels spectra were acquired in a vg 501 scanning transmission electron microscope , provided with a liquid nitrogen cooling system at the sample stage ( 150 k ) .
low temperatures have been proven to reduce the mass loss damage by decreasing the atomic mobility.@xcite transmitted electrons were collected on a liquid nitrogen cooled ccd camera with a low read - out noise of three counts r.m.s . and a negligible dark count noise
. the energy dispersion of the spectra was 0.27 ev , to allow simultaneous acquisition of carbon and oxygen k - edges .
the effective energy resolution , resulting from microscope instabilities and spectrometer aberrations , is estimated to 0.5 ev on the carbon k - edge .
the accelerating voltage has been limited to 60 kv , and a low electron dose working mode was set by optimization of the acquiring parameters ( beam current , illumination area , and time ) on ( r)go few layer flakes . during spectrum imaging
the beam focus ( i.e. , illumination area at the sample stage ) has been adjusted to match the pixel size in order to avoid oversampling .
the exposure of the material to the electron beam before spectra acquisition has been prevented by a fast blanking system before the sample , following the same procedure as in ref .
equally , eels spectra were acquired prior to the corresponding stem images .
elemental quantification was performed considering carbon and oxygen edges , within a 25 ev energy window .
specific cross sections were derived from the hartree - slater model .
an error of a few percent may be assumed on the quantification results.@xcite eels spectra collected in a low electron dose mode are characterized by little intensity of the signals .
oxygen quantification has been performed on denoised spectra .
principal component analysis ( pca ) represents a common technique for noise reduction in core eels elemental mapping .
@xcite pca analysis has been performed on spectrum images using the hyperspy software.@xcite after remarking significant changes in carbon fine structure peaks , the number of pca components ( less than 10 in the usual routine ) was raised to 15 .
fine structure analysis has been performed on the as - acquired spectrum images , not treated with pca , in order to avoid the introduction of any artifacts .
nevertheless , in order to appreciate fine structure peaks , a far more intense signal with respect to quantification is required .
the signal intensity has been improved by summing spectra arising from contiguous areas up to few tens of nanometers wide ( pixels in the spectrum image ) displaying a uniform oxygen content .
[ [ dft - simulations . ] ] dft simulations .
+ + + + + + + + + + + + + + + + structural optimizations have been performed within the framework of density functional theory ( dft ) under the local density approximation ( lda ) as implemented in the aimpro code.@xcite carbon , oxygen and hydrogen pseudopotentials are generated using the hartwingsen - goedecker - hutter scheme.@xcite valence orbitals are represented by a set of cartesian gaussians of s- , p- , and d - type basis functions centered at the atomic sites .
we have used a basis set of 22 independent functions for carbon ( pdpp gaussian exponents ) , 40 functions for oxygen ( dddd ) , and 12 functions for hydrogen ( ppp ) .
the optimization of the cell parameter was performed simultaneously with the relaxation of the single atomic positions , using a conjugate gradients algorithm .
[ [ optimizing - the - dose . ] ] optimizing the dose .
+ + + + + + + + + + + + + + + + + + + + in order to define optimal conditions for gentle illumination with a 60 kev electron beam , we have monitored the evolution of oxygen content and carbon near - edge fine structures in go as a function of the irradiating electron dose . in fig.[chronospim ] we present a time series of core eels spectra acquired under homogeneous illumination over an extended area of a go flake ( @xmath4 nm@xmath5 ) with a 4 pa electron beam current and an integration time of 2.4 s , corresponding to an electron dose rate as low as @xmath6 e@xmath7@xmath8s@xmath9 .
the illuminated region has an initial oxygen content of about 45% ( as derived from the spectrum in fig.[chronospim].a ) that progressively reduces as a result of the increasing electron dose , as shown in fig.[chronospim].b .
furthermore , the lowering of the oxygen content is associated with significant changes in c k - edge elnes fine structures ( fig.[chronospim].c ) .
variations in the near - edge fine structures are a typical indicator of radiation damages affecting chemical bonding and atomic orbital hybridization.@xcite on this basis , it has been possible to define an upper limit for the electron dose of about @xmath10 e@xmath11@xmath8 , below which no substantial changes occur neither in the flakes stoichiometry nor in eels fine structures .
the need of low electron dose imposes constraints on stem spatial resolution.@xcite indeed , the total acquired spectral intensity is proportional to the electron dose ( fixed by irradiation conditions ) , the elemental specific cross section , the surface projected density of atoms i.e. the number of scattering centers ( depending on the sample thickness ) and the size of the illuminated area .
gentle illumination lowers the intrinsically weak intensity of core eels signals and high spatial resolution implies that a very low number of scattering centers contribute to the eels signal . in the specific setup of our microscope ,
the highly efficient ccd camera and the higher electron dose granted by sample cooling@xcite allow for improving the eels signal intensity .
the signal / noise ( s / n ) ratio that is compatible with elemental quantification ultimately defines a @xmath13 nm wide minimal illumination area for the study of go . [
[ elemental - quantification . ] ] elemental quantification .
+ + + + + + + + + + + + + + + + + + + + + + + + + ] the spatial distribution of oxygen in go and rgo has been investigated by stem - eels spectrum imaging .
this acquisition mode consists in collecting simultaneously the haadf intensity and a complete eels spectrum at each position of the electron beam during the scan , with a @xmath13 nm minimal beam size and scanning step ( no oversampling ) , to be compatible with the illumination conditions defined above . in fig.[carta - go ]
we report an example of eels oxygen quantification on an individual few layer go flake .
the haadf and bright field ( bf ) images of the whole flake are presented in fig.[carta - go].a - b , respectively .
a spectrum image has been acquired over the @xmath12 @xmath13 wide area marked in fig.[carta - go].a and the resulting haadf image is shown in fig.[carta - go].c .
acquisition parameters were 10 nm spatial resolution ( probe size and scanning step ) , a 15 pa current and 0.15 s integration time , that correspond to an @xmath1@xmath14 e@xmath11@xmath8 total electron dose per spectrum .
the haadf intensity depends on both the amount of matter ( thickness and areal density ) and the chemical species ( atomic number ) .
this information appears separated in the integrated k - shell ionization edges maps derived from the spectrum image ( fig.[carta - go].d - e ) , that show the spatial variation of the surface projected amount of carbon ( hence the number of carbon layers ) and oxygen , respectively . in the quantification map of oxygen relative to carbon ( fig.[carta - go].f )
two well separated domains with oxidation levels centered at @xmath132 at.% and @xmath140 at.% are visible , as confirmed by the associated image histogram in fig.[carta - go].g .
a third less extended domain ( bottom left corner in the map ) corresponds to @xmath145 at.% with maximum local content of almost 50 at.% . when interpreting these results it must be considered that for each pixel the eels signal is integrated over the specimen thickness .
hence , a nonhomogeneity in the relative quantification map can result from a spatial variation of the local oxidation level in regions with a constant number of layers or from the local superposition of additional layers with a different oxidation level . in fig.[carta - go].f
an increase of the oxygen relative content is visible from the left to the right side of the map forming two well separated regions .
the oxygen integrated k - edge map ( fig.[carta - go].e ) shows the same trend : the right side area does contain a higher amount of oxygen . on the contrary ,
the intensity of the carbon integrated k - edge ( fig.[carta - go].d ) is almost uniform , indicating that the flake has a constant carbon thickness along the region displaying the oxygen gradient .
therefore these results show a variation of the oxidation level within an individual flake , forming spatially well separated oxidation phases which are not correlated with the flake thickness .
overall the oxygen content in go has been observed to vary between @xmath110 at.% and @xmath150 at.% ( c / o ratio of 9:1 and 1:1 , respectively ) . the higher oxidation level ( 30 - 50 at.% ) seems to characterize about one - quarter of the analyzed material , while in extremely rare cases the oxygen content falls below the detection limit ( unoxidized flakes ) , confirming the heterogeneity of go .
therefore despite the fact that the local oxygen concentration can be extremely high , the average oxygen content in individual flakes is @xmath125 at.% , which is in good agreement with spatially averaged values reported in the literature.@xcite regions up to a few hundred nanometers wide are characterized by an almost uniform oxidation level ( at a 3 - 10 nm scale ) , while the transition between two oxidation phases occurs in the space of few tens of nanometers .
this observation is compatible with a recently proposed oxidation mechanism where the reaction progresses within graphite flakes in a front - like diffusive - controlled pathway.@xcite ] stem images and spectroscopic analysis from rgo are shown in fig.[carta - rgo ] .
haadf and bf images of the whole flake are shown in fig.[carta - rgo].a - b .
a spectrum image has been collected over the region marked in fig.[carta - rgo].a , over an area of about @xmath15 nm@xmath5 .
acquisition parameters were as follows : 3 nm spatial resolution , 11 pa current , and 0.04 s integration time , for a total electron dose per spectrum of @xmath1@xmath16 e@xmath11@xmath8 .
fig.[carta - rgo].c - g shows the associated haadf image , the integrated k - shell ionization edges maps of carbon and oxygen , the quantification map of oxygen relative to carbon , and its associated image histogram . as in go , the carbon and oxygen k - edge
integrated intensity maps appear mainly noncorrelated .
weak correlations appear occasionally , not systematically , in some limited areas .
namely , in the central region of the spectrum image the carbon k - edge intensity is almost constant , while the oxygen k - edge signal varies and may be responsible for the contrast variation in the haadf image .
the distribution of oxygen forms patches few tens of nanometers wide ( fig.[carta - rgo].e - f ) that correspond to an unimodal distribution centered at about 5 at.% ( fig.[carta - rgo].g ) , with a local upper limit of 12 at.% ( local residual oxygen up to 20 at.% has been observed in other flakes , over limited regions ) .
these values are situated in the range of spatially averaged oxygen quantifications reported in the literature.@xcite [ [ core - eels - fine - structures . ] ] core eels fine structures .
+ + + + + + + + + + + + + + + + + + + + + + + + + + ] besides oxygen quantification , coordination states and chemical bonds for different oxidation phases can be extracted from spectrum images through energy - loss near - edge structures analysis .
the spatial resolution of the spectrum images has been optimized for elemental analysis ; the higher s / n ratio required for fine structure investigation can be obtained by integrating the spectrum images over regions few tens of nanometers wide , having a uniform oxygen content . in fig.[finestructure - c ] we present fine structures at the carbon k - edge that are representative of particular ranges of oxygen content .
fig.[finestructure - c].a shows the fine structure signatures observed in go areas characterized by 10 to 50 oxygen at.% . in both spectra corresponding to low oxygen concentration ( @xmath120 at.% ) , the @xmath17 and @xmath18 signatures associated with sp@xmath5 carbon are visible respectively at 285.0 ev and 292.7 ev.@xcite two additional peaks at 287.2 ( a ) and 288.2 ev ( b ) could be isolated over two distinct areas of go ( in this particular case showing 20 and 25 oxygen at.% , respectively ) .
these features generally show a feeble intensity and can appear simultaneously in the same spectrum , resulting in mixed features that can hardly be identified . in go regions containing @xmath130 oxygen at.% ,
the intensity of the @xmath17 peak lowers , while the 287.2 ev peak ( a ) remains intense ( 32 oxygen at.% for the here reported spectrum ) .
an additional feature rises at 290.1 ev ( c ) followed by a broad triangular - shaped signal at about 300 ev ( d ) . for
higher oxidized go regions ( 40 - 50 at.% , the reported spectrum corresponding to 40% ) , the @xmath17 peak further lowers , while the peaks at 287.2 ev ( a ) and 290.1 ev ( c ) get stronger and very well - defined .
the 290.1 ev peak appears particularly sensitive to irradiation damages , as shown by fig.[chronospim].c .
after reduction ( fig.[finestructure - c].b ) the sharpness of the @xmath17 and @xmath19 peaks , with a trace of @xmath19 excitonic feature at 291.6 ev @xcite , reveals a very good recovery of the graphitic network .
an additional well - defined peak at 287.2 ev ( a ) is observed ( the two spectra shown correspond to 7 oxygen at.% and 12 at.% ) . rarely and in some limited regions , other weak fine structure peaks have been distinguished , i.e. peak b ( 5 oxygen at.% ) .
these a and b spectral features are below the detection limit for lower oxygen concentrations .
these results show that specific near - edge fine structures and thus coordination states of the carbon atoms correspond to different oxidation rates .
the energies of eels peaks in fig.[finestructure - c ] are compatible with previous xanes results which however could not separate the specific near - edge fine structures shown here due to spatial averaging over wide regions with different oxygen levels .
absorption peaks indicated as a - c had been attributed in the literature to c - o @xmath20 symmetry antibonds : lower energy peaks to single bonds and higher energy peaks to double bonds .
more precisely , hydroxyl groups had been generally associated with peak a , epoxide groups to peak a or b , and carbonyl and carboxyl groups to peak b or c.@xcite in the @xmath19 region , carbonyl groups have been related to peak d.@xcite in a recent stem - eels investigation a shallow peak a has been observed at the carbon k - edge and assigned to epoxide , while no peak c has been detected.@xcite these assignments , still largely debated , have been obtained by comparison with reference spectroscopic signatures of aromatic molecules@xcite or spectra simulations on hypothetical go atomic structures.@xcite however , as shown a number of times by numerical simulations , @xcite the spatial configuration of functional groups within ( r)go deeply affects their orbital associated energies , and these might strongly differ from those of the same groups within molecules . moreover , the sensitivity of the material suggests the need of caution while comparing spectroscopic results .
in particular , the illumination conditions can induce changes in the relative intensity of fine structure peaks by transformation of the functional groups ( possibly hydroxyls into epoxides ) or selective sputtering of the oxygen groups .
[ [ structural - model . ] ] structural model .
+ + + + + + + + + + + + + + + + + the structure of go can be revisited on the basis of the here reported spatially resolved spectroscopic results .
several complex atomic models for go have been proposed in the last 20 years .
@xcite oxygen adsorption is usually described in terms of the basic functional groups of epoxide , hydroxyl , carbonyl ( -co ) and carboxyl ( -cooh ) .
the last two groups involve multiple bonding with a single carbon atom , which can occur at edges or at defects within the graphene lattice . on the contrary
, a complete and in principle reversible functionalization of a perfect graphene network can be achieved with epoxides and hydroxyls , resulting in a c / o ratio of 2:1 ( 33 oxygen at.% ) for solely epoxides functionalization , 1:1 ( 50 oxygen at.% ) for solely hydroxyls functionalization , and intermediate values for mixed structures . up to now , the lerf - klinowski model@xcite has been the most referred to . in this representation , different oxygen groups are randomly distributed on both sides of the carbon sheets , with coexisting graphitic and oxidized regions , thus forming a partially functionalized structure .
hydroxyls and epoxides are supposed to be the dominant groups on the basal plane , while carboxyls are expected to fill carbon dangling bonds at the sheet edges and vacancy sites.@xcite carbonyl groups as ketones and/or quinones are also generally expected at the plane edges . a high proximity among hydroxyls and epoxides and the presence of five- and six - membered lactols and ester carbonyls
have been proposed on the basis of nmr studies.@xcite with respect to our experimental observation , the lerf - klinowski model is consistent with go regions with a 20 oxygen at.% content , but can not account for the much higher oxygen concentration locally observed by eels ( @xmath145 at.% ) .
the presence of water molecules in the atomic structure of go is still under debate .
water molecules are expected to form hydrogen bonds with oxygen functional groups on go.@xcite we expect reasonably that any possible intercalated water would have been removed from the here observed flakes , during the preparation process .
indeed , differently from the common protocol where go is directly deposited on a substrate as a water dispersion , in this work we employed dried go which was successively redispersed in ethanol and ultrasonicated .
moreover observation conditions involve high vacuum levels ( @xmath3 mbar ) , resulting in evaporation of residual water molecules . on the basis of these considerations
we expect a negligible amount of water .
this hypothesis is corroborated by our observed eels fine structures .
indeed , the strong intensity and sharpness of the fine structure peaks at the carbon k - edge correlated with a high oxygen content ( fig.[finestructure - c].a ) suggest that a significant amount of carbon atoms is involved in a specific kind of carbon - oxygen bond .
thus , the @xmath145 oxygen at.% in go can not arise from intercalated water and should be ascribed to oxygen directly bound to carbon .
recently it has also been proposed that the go surface could be decorated by strongly bound highly oxidized debris coming from the oxidation process .
@xcite in the present work , an important contribution of such debris to the calculated oxygen content is not compatible with our spectroscopic investigations : their presence would result in an increase of intensity in the k - edge maps of both oxygen and carbon ( as an additional go layer ) and hence in a correlation between the two maps , which is not observed in fig.[carta - go].d - e .
moreover , the highly oxidized residues could contain at most @xmath150 oxygen at.% , considering a fully functionalized structure ( hydroxyls on the basal plane and carbonyls - carboxyls at the edges , as discussed above ) . in fig.[carta - go].g
the local oxygen content reaches almost 50 at.% and represents an integrated measurement over the flake thickness : even assuming the presence of debris , the underlying go layer(s ) has to be necessarily almost 50 at.% oxidized .
this experimental evidence reinforces the criticism recently raised against the debris go model .
@xcite first principle simulations have shown that in go there is a strong driving force toward a phase separation in fully functionalized regions ( epoxides and/or hydroxyls ) and pristine graphene regions.@xcite the limit of 50 at.% local oxygen concentration observed in this work is not consistent with a full functionalization dominated by epoxides , since the oxygen content would tend to 33 at.% . a large presence of carboxyl and carbonyl groups
should also be excluded , because such a highly defective structure could hardly be reduced to a highly graphitic arrangement compatible with the carbon k - edge fine structures observed in rgo flakes .
go containing almost 50 at.% of oxygen and converting into a highly graphitic structure when reduced can be explained only by an almost hydroxyl saturated graphene lattice .
this limit structure is analogous to graphane@xcite ( i.e. , hydrogen saturated graphene ) , and hydroxyl groups should alternate at the two sides of the carbon plane in order to minimize strain.@xcite this limit configuration is presented in fig.[model ] after optimization by dft - lda .
the graphene network is shown to be strongly distorted , with carbon atoms puckered out of plane by about 0.26 leading to a change of their hybridization state from pure sp@xmath5 to partial sp@xmath0 .
in contrast with graphane , where the sp@xmath0 carbon hybridization reduces the in plane cell parameter ( @xmath12.42 ),@xcite we obtain an increased cell parameter of 2.58 , ( about 5% higher than dft - lda optimized graphene ) and a c - c bond length higher than diamond ( 1.59 ) . oxygen steric hindrance and hydrogen bonding between neighboring groups forming hydroxyl chains are responsible for this in plane expansion as for graphene edges functionalized by hydroxyl groups where the accumulated strain is released via the formation of static out - of - plane ripples.@xcite it might be challenging to directly observe this lattice expansion through diffraction because uniform and continuous highly functionalized domains are of little extension .
furthermore , the accurate determination of the unit cell parameter by electron diffraction is a nontrivial technique itself , for in 2d materials the reciprocal space of monolayers consists of rods and thus the total intensity of the diffraction spots changes little as a function of the tilt angle , while the spots broadening is important .
@xcite eels fine structures for highly oxidized go ( 30 - 50 oxygen at.% , fig.[finestructure - c].a ) can thus be interpreted in the limit of this hydroxyl saturated graphene model .
the extremely low intensity of the @xmath17 peak can be ascribed to the loss of sp@xmath5 hybridization at carbon atoms .
the 287.2 ev peak ( a ) , which is observed also for lower oxygen rates , can be assigned to c - oh antibonds , in agreement with previous xanes works.@xcite the peak at 290.1 ev ( c ) can not be associated with c - o antibonds of other oxygen functional groups , because they are expected to be present only in a small amount and this is not compatible with the strong intensity of the peak .
nevertheless , the interpretation of this feature remains unclear .
fine structures for lower oxidation rates ( 10 - 30 oxygen at.% , fig.[finestructure - c].a ) are much less defined .
this can be due to a higher variety of functional groups and the high number of possible configurations that they can assume at the graphene network .
the presence of the @xmath17 peak and of a weak sp@xmath5 @xmath19 peak infers a limited presence of nonfunctionalized carbons . aside from peak
a ascribed to hydroxyl groups , peak b can be attributed to epoxide groups .
after reduction the very intense and sharp peak a observed at higher residual oxygen regions ( fig.[finestructure - c].c , @xmath115 oxygen at.% ) would then correspond to the presence of solely hydroxyl groups .
this interpretation is compatible with the identification of several simple reaction routes for epoxide reduction by hydrazine , while hydroxyl removal mechanisms are still debated.@xcite considering the same assignments as in go , the weak peak b in rgo indicates a small residual amount of epoxides .
again , the difference in the intensity of features a and b can derive from their relative amount ( hydroxyl probably remain the dominant species also after reduction ) and the different number of configurations that these functional groups can form at the carbon network .
in conclusion we have investigated the oxygen distribution in go and rgo by core eels spectroscopy in a stem microscope .
a spatial resolution of 3 nm for spectrum images has been achieved by reducing radiation damage through optimized acquisition conditions and an experimental setup combining a low noise ccd detection camera with a liquid nitrogen cooled sample stage .
previous spectroscopic quantifications have provided only overall oxygen contents , and go and rgo have been often described as chemically homogeneous materials at a nanometric scale .
we have shown that within individual flakes the oxygen content strongly varies , forming homogeneous domains with a lateral side of few tens of nanometers . in go different oxidation levels have been identified , with a maximum local concentration of almost 50 at.% of oxygen , corresponding to a c / o ratio of 1:1 . in rgo
the residual oxygen mainly concentrates in domains with about 5 oxygen at.% .
specific fine structures at the carbon k - edge can be clearly associated with different oxidation phases .
highly oxidized go displays a shallow graphitic signature with extremely intense and sharp additional peaks , not observed before . in rgo
the graphitic character is extensively recovered with weakened additional features .
the assignment of the spectroscopic peaks is still debated in the literature , and the previously proposed mixed epoxides - hydroxyls atomic model for go can not account for the highest oxygen concentrations observed in this work .
we propose a limit structural model for these highly oxidized regions , in which all carbon atoms are functionalized with -oh groups , leading to an sp@xmath0 carbon 2d material analogous to graphane . a strong signature for -oh antibonds
can then be identified in go carbon near - edge fine structures . in rgo , a sharp -oh antibond peak , localized at
highest oxidized regions , infers that this group is the dominant surviving species .
the presence of domains fully covered with hydroxyl groups has critical implications for the engineering of the electronic structure of individual flakes .
indeed , it strongly affects the hydrophilicity of the flakes and the interaction with molecules from the gas and vapor phase.@xcite furthermore , the presence of a dense hydroxyl layer on ( r)go sheets has important consequences for the anchoring of metal and ceramic nanoparticles or biomolecules and the assembly of individual flakes into corresponding macroscopic forms.@xcite
a.t . , a.z . and o.s .
acknowledge support from the agence nationale de la recherche ( anr ) , program of future investment tempos - chromatem ( no .
anr-10-eqpx-50 ) .
the work has also received funding from the european union in seventh framework programme ( no .
fp7/2007 -2013 ) under grant agreement no . n312483 ( esteem2 ) .
a.m.b . and w.k.m .
are grateful for financial support from the spanish ministry mineco and the european regional development fund ( project ene2013 - 48816-c5 - 5-r ) and from the regional government of aragon and the european social fund ( dga - esf - t66 grupo consolidado ) .
the authors are grateful to p. launois , s. rouziere , and c. p. ewels for useful discussions .
62ifxundefined [ 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 link:\doibase 10.1002/adma.200903689 [ * * , ( ) ] link:\doibase 10.1039/b917103 g [ * * , ( ) ] link:\doibase 10.1038/nchem.907 [ * * , ( ) ] link:\doibase 10.1039/c2ra00663d [ * * , ( ) ] link:\doibase 10.2307/108699 [ * * , ( ) ] link:\doibase 10.1002/cber.18980310237 [ * * , ( ) ] link:\doibase 10.1021/ja01539a017 [ * * , ( ) ] link:\doibase 10.1016/j.carbon.2011.11.010 [ * * , ( ) ] link:\doibase 10.1021/jp9731821 [ * * , ( ) ] link:\doibase 10.1038/nmat3316 [ * * , ( ) ] link:\doibase 10.1038/srep02484 [ * * , ( ) ] link:\doibase 10.1021/cm203223z [ * * , ( ) ] link:\doibase 10.1021/nn3047378 [ * * , ( ) ] link:\doibase 10.1021/nn900694 t [ * * , ( ) ] link:\doibase 10.1002/adma.201000732 [ * * , ( ) ] link:\doibase 10.1021/nl9031617 [ * * , ( ) ] link:\doibase 10.1016/j.carbon.2010.09.063 [ * * , ( ) ] @noop _ _ , ed .
( , , ) link:\doibase 10.1021/nl8034256 [ * * , ( ) ] link:\doibase 10.1016/j.carbon.2015.06.025 [ * * , ( ) ] link:\doibase 10.1016/j.ultramic.2012.07.006 [ * * , ( ) ] link:\doibase 10.1016/j.carbon.2011.09.042 [ * * , ( ) ] link:\doibase 10.1016/j.carbon.2007.02.034 [ * * , ( ) ] link:\doibase 10.1016/s0304 - 3991(80)80009-x [ * * , ( ) ] link:\doibase 10.1038/nnano.2010.105 [ * * , ( ) ] link:\doibase 10.1016/j.ultramic.2008.01.004 [ * * , ( ) ] link:\doibase 10.1016/j.ultramic.2010.10.001 [ * * , ( ) ] link:\doibase 10.5281/zenodo.16850 [ `` , '' ] ( ) in http://www.sciencedirect.com/science/article/pii/s0080878408630586[__ ] , , vol . , ( , ) pp . link:\doibase 10.1016/j.cpc.2007.08.007 [ * * , ( ) ] link:\doibase 10.1103/physrevb.58.3641 [ * * , ( ) ] link:\doibase 10.1021/nn500606a [ * * , ( ) ] link:\doibase 10.1088/1742 - 6596/26/1/035 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.71.3725 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.74.614 [ * * , ( ) ] link:\doibase 10.1209/0295 - 5075/82/67004 [ * * , ( ) ] link:\doibase 10.1021/jp203741y [ * * , ( ) ] link:\doibase 10.1016/j.carbon.2010.12.002 [ * * , ( ) ] link:\doibase 10.1039/c1jm11071c [ * * , ( ) ] link:\doibase 10.1021/jp306497f [ * * , ( ) ] link:\doibase 10.1038/srep02250 [ * * , ( ) ] link:\doibase 10.1038/srep04525 [ * * , ( ) ] link:\doibase 10.1021/j100195a018 [ * * , ( ) ] @noop * * , ( ) link:\doibase 10.1002/adma.201401300 [ * * , ( ) ] link:\doibase 10.1021/ja8021686 [ * * , ( ) ] link:\doibase 10.1103/physrevb.79.125435 [ * * , ( ) ] link:\doibase 10.1021/jp108688 m [ * * , ( ) ] link:\doibase 10.1103/physrevb.82.161406 [ * * , ( ) ] link:\doibase 10.1021/jp204476q [ * * , ( ) ] link:\doibase 10.1021/ja9030243 [ * * , ( ) ] link:\doibase 10.1038/nchem.281 [ * * , ( ) ] link:\doibase 10.1002/anie.201007520 [ * * , ( ) ] link:\doibase 10.1016/j.carbon.2015.05.058 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.103.086802 [ * * , ( ) ] link:\doibase 10.1126/science.1167130 [ * * , ( ) ] link:\doibase 10.1103/physrevb.82.125403 [ * * , ( ) ] link:\doibase 10.1103/physrevb.84.134110 [ * * , ( ) ] link:\doibase 10.1038/nature05545 [ * * , ( ) ] link:\doibase 10.1021/jp909284 g [ * * , ( ) ] link:\doibase 10.1021/nl5013689 [ * * , ( ) ] link:\doibase 10.1016/j.carbon.2014.08.020 [ * * , ( ) ] | the type and distribution of oxygen functional groups in graphene oxide and reduced graphene oxide remain still a subject of great debate .
local analytic techniques are required to access the chemistry of these materials at a nanometric scale .
electron energy loss spectroscopy in a scanning transmission electron microscope can provide the suitable resolution , but go and rgo are extremely sensitive to electron irradiation . in this work
we employ a dedicated experimental setup to reduce electron illumination below damage limit .
go oxygen maps obtained at a few nanometers scale show separated domains with different oxidation levels .
the c / o ratio varies from about 4:1 to 1:1 , the latter corresponding to a complete functionalization of the graphene flakes . in rgo
the residual oxygen concentrates mostly in regions few tens nanometers wide .
specific energy - loss near - edge structures are observed for different oxidation levels . by combining these findings with first - principles simulations we propose a model for the highly oxidized domains where graphene is fully functionalized by hydroxyl groups forming a 2d - sp@xmath0 carbon network analogous to that of graphane . |
since the first observation of a transiting exoplanet @xcite , knowledge of exoplanetary radii , composition and atmospheres has grown explosively . in - transit and out - of - transit spectroscopy and photometry have revealed water absorption in hd 189733b @xcite , atmospheric emission in tres-1 @xcite , a surprising number of anomalously large planets @xcite , and constraints on exoplanetary composition @xcite .
accurate transit timing can also indicate the presence of additional bodies in the system through perturbations to the transiting planet s orbit @xcite . when transits are observed in stellar lyman-@xmath4 , @xmath5 , @xmath6 , @xmath7 and @xmath8 emission , they show much deeper minima than for visible wavelengths , revealing escaping atmospheres extending far beyond the geometric radii of planets @xcite
. these light curves also constrain atmospheric conditions and mass escape from the planet s roche lobe @xcite .
transit searches have been largely in the optical and near - infrared continuum , where the star is optically thick and limb darkened due to the temperature profile at the @xmath9 1 surface of the star . for limb darkened wavelengths , the flux from the transiting system is at a minimum when the planet crosses the sub - earth longitude of the star ( phase = 0.0 ) because the stellar disk is brightest at its center .
by contrast , emission lines from stellar chromospheres and transition regions can be limb _
brightened_. for optically thick emission , limb brightening occurs when the source function increases with stellar altitude . for optically thin emission ,
strong limb brightening occurs because the chromospheric and transition region gas has its largest column density at the edges .
@xcite showed that transit observations of chromospheric emission lines decrease sharply to a minimum at the first limb , increases to a local maximum mid - transit and then reverses the process as the planet exits the stellar disk .
as a consequence , such a transit curve will be `` w''-shaped .
@xcite point out that limb brightening could be useful for detecting exoplanetary transits of giant stars
. light curves of limb brightened wavelengths have deeper minima than for both limb darkened and uniform disk emission .
the star emits over a smaller effective area a ring instead of a disk so the planet covers a larger amount of the stellar flux .
this is important for transits of giant stars where broadband transit depths can be below 0.01% for jupiter - sized planets .
also , the planet covers its host s limb for a small fraction of the transit , allowing for feasible detection of giant star transits with ground based telescopes , which suffer from systematic photometric errors over timescales longer than one night .
the exoplanets 4uma b , hd 122430b , hd 13189b , and hip75458 b all have transit probabilities greater than 10% and may be useful targets for future studies @xcite .
@xcite approximate the limb brightened star as a central disk of emission surrounded by a circularly symmetric ring with @xmath1030 times the intensity . with this ring approximation ,
the maximum depth of the transit is proportional to the ratio of the planet radius to stellar radius , @xmath11 if the emission from the central disk is negligible , instead of @xmath12 , as expected for a uniform disk .
this is because the planet covers @xmath102r@xmath13 out of a circle of emission whose total circumference is 2@xmath14 . in this paper
, we present a transit light curve calculation for an optically - thin and geometrically - thin shell of emission .
the calculated maximum transit depth scales as @xmath15 , instead of @xmath11 . in @xmath16[labl : chromlcurve ] we calculate the expected limb brightened light curve for an optically thin emission line .
we consider optically thin emission because it shows at least 8 times the limb brightening of optically thick emission @xcite .
we fit this model light curve to @xmath0 emission from hd209458 in @xmath16[osiris ] and discuss the implications for hd209458b s thermosphere in @xmath16[discuss ] .
[ labl : thinshell ] we make the approximation that the thickness of the chromosphere , @xmath17 , is much smaller than the size of the planet ( @xmath18 ) and star ( @xmath19 ) . under this geometrically thin approximation ,
the total flux from the star is proportional to the surface area of the hemisphere facing the earth times its thickness @xmath17 , because the total flux for an optically thin emission line is proportional to the total number of emitting ions . neglecting any photospheric contribution , the amount of emission that the planet blocks is then simply the amount of the stellar surface that the planet covers times @xmath17 . in this geometrical limit ,
therefore , the thickness of the chromosphere , @xmath17 , cancels out .
emission , whose transit light curve is given by equation [ disk].,scaledwidth=50.0% ] we compute the light curve for zero thickness ( @xmath20 ) by finding the area of the the planet s shadow and dividing this by the surface area of a hemisphere with radius @xmath19 .
the volume of the intersection of a cylinder with a sphere is given by @xcite and we find the surface area of the intersection by taking a partial derivative with respect to the radius of the sphere .
the result can be expressed analytically in terms of elliptic integrals .
let @xmath21 be the distance , in units of stellar radii , @xmath19 , from the center of the planet to the center of the star projected onto a plane perpendicular to the observer .
let @xmath22 be the planet / star radius ratio , r@xmath13/r@xmath2 . to calculate the light curve for a planet that does not pass through the center of the star
, one can simply write @xmath23 as @xmath24 , where @xmath25 is the distance to the closest approach point in stellar radii @xmath19 and @xmath26 is the impact parameter in stellar radii @xmath19 .
the transit depth , @xmath27 is a piecewise function with three different regimes : ( 1 ) when the planet is fully contained in the stellar disk ( 2 ) when the planet is at egress / ingress and ( 3 ) when the planet is beyond the stellar disk .
we also include the case that the planet s absorption profile is larger than the star . in this case , the same formulae apply
see equation [ defs ] . a small figure for each regime is included in the equations , where an open circle represents the star and a filled circle represents the planet .
@xmath28 \nonumber\end{aligned}\ ] ] @xmath29 where @xmath30 , @xmath31 , and @xmath32 are the complete legendre elliptic integrals of the first , second , and third kinds , and @xmath33 is the heaviside step function . for the elliptic integrals , we use the conventions of @xcite where @xmath34 for the third elliptic integral . is located at ` http://www.astro.washington.edu/agol/ ` . ]
these formulae are difficult to evaluate numerically at @xmath35 , @xmath36 and @xmath37 due to the formal divergence of different terms in equations [ disk ] and [ defs ] ; the divergences cancel out analytically , but routines that evaluate the elliptic integrals diverge . however , these locations are a set of measure zero , and thus are tractable when modeling data .
the transit is deepest slightly before second contact , so we can estimate the maximum transit depth as follows from figure [ fig01 ] by comparing the total emitting area of the star to the total stellar surface area blocked by the planet .
the blocked surface is the same as the shadow produced by a sphere of radius @xmath18 onto a hemisphere of radius @xmath19 . when the planet occults the edge of the star ( second contact ) , as shown from an edge - on viewpoint in this figure , then the length of the arc of the long axis of the shadow is @xmath38 .
the diameter of the planet is @xmath39 , where the latter approximation is valid for @xmath40 .
we can approximate the shadow as an ellipse with a semi - minor axis of @xmath18 and a semi - major axis of @xmath41 , so the area of the shadow is @xmath42 .
thus , the maximum depth of transit is given by @xmath43 which is accurate to within 5% for r@xmath13/r@xmath2@xmath44 0.23 .
note that this is a different scaling for maximum transit depth than that given in @xcite who assume that most of the stellar emission is from a thin ring . for r@xmath13/r@xmath2 = @xmath45 ,
half of the stellar emission is within @xmath18 of the stellar radius @xmath19 , so for r@xmath13/r@xmath2 @xmath46 , the ring approximation is valid , but the scaling of transit depth assumes negligible limb curvature at the scale of the planet ( r@xmath13/r@xmath2 @xmath47 1 ) .
the remarkable consequence of equation [ depth_anal ] is that the depth of a chromospheric transit does not decline as much with the radius of the planet as a transit of a uniform disk .
a chromospheric transit has a maximum depth that is @xmath48 times deeper than the maximum transit depth of a uniform disk ; thus smaller planets have an advantage to be observed at chromospheric wavelengths , as emphasized by @xcite .
it should also be noted that at mid - transit , the double - u curve has a smaller transit depth than for a uniform disk , because the planet covers only @xmath49 out of a hemisphere of area @xmath50 .
figure 3 ( solid curve ) shows the transit light curve , 1-@xmath27 , for a planet that has r@xmath13/r@xmath2=0.08 .
the estimate given by equation [ depth_anal ] and the more detailed equation [ disk ] agree well , predicting maximum chromospheric depths to be 1.8 times deeper than for uniform disk brightness .
( for a uniform disk emission @xmath51=@xmath52 ) .
we also include a light curve for the thin circle of emission shown in figure [ limbmodel ] b for comparison .
[ fig02 ] emission from the solar limb as observed by sumer ( solar ultraviolet measurement of emitted radiation ) @xcite is strongly limb - brightened , indicating that a transit of this emission line should have a double - u light curve .
( b ) @xmath53 304 image of the solar limb , taken by the extreme - ultraviolet imaging telescope @xcite is , by contrast , not limb brightened because it is optically thick in the chromosphere and thus the additional column density at the limb does not contribute any more flux than the central disk.,scaledwidth=50.0% ]
@xcite observed the exoplanet host hd 209458 with the hubble space telescope imaging spectrograph ( stis ) instrument and found an extended hydrogen , oxygen and carbon atmosphere around the planet by fitting the light curves to the @xmath54 , @xmath8 and @xmath5 lines during a transit .
@xcite found no @xmath0 absorption when fitting the light curve to a model of a spherical planet occulting a _ uniform _ stellar disk .
we fit their @xmath0 1394 data ( @xcite , figure 3 ) with a double - u model given by equation [ disk ] .
this model is appropriate for the @xmath0 emission , as evident in the solar image in figure [ limbs ] ( a ) where strong limb - brightening is apparent .
@xcite found that the @xmath0 1394 / @xmath0 1403 ratio is 2:1 within the errors , indicating that the @xmath0 1394 emission is optically thin @xcite .
we investigated the transit of the @xmath0 line because it the strongest optically thin line in hd209458b s stis spectrum . while other limb brightened emission lines
do exist , we focus on the optically thin ones because they should be the most limb brightened .
the model has two free parameters : the planet / star radius ratio r@xmath13/r@xmath2 and an overall constant that sets the off - transit flux .
the second parameter is necessary since the off - transit @xmath0 1394 stis flux is poorly constrained .
the impact parameter is fixed with a value of @xmath55 , using an inclination of 86.7@xmath56 and a semi - major axis @xmath57 @xcite .
figure 5 shows the light curve for the @xcite data and the best fit model .
in addition to the best fit model , we show two more limb brightened models where r@xmath13/r@xmath2 is a fixed parameter for comparison
. these have r@xmath13/r@xmath2 equal to 0 and 0.12 , representing no @xmath0 absorption and the geometric planet radius @xcite , respectively . using a levenberg - marquardt fitting algorithm @xcite and the points at which the @xmath58 = 1 to calculate uncertainties , we find that r@xmath13/r@xmath2 = 0.34@xmath3 , close to the size of the roche lobe .
the double - u model has a total @xmath59 of 10.4 with 12 data points and two parameters , which is 3.1 less than fitting the data to a constant flux line ( representing no detection ) with one parameter .
since these models are nested , we can use the maximum likelihood ratio test .
the p - value for the difference in @xmath59 is 0.05 , so the data favor the double - u model with 95% confidence , assuming normally distributed data .
the same test favors the r@xmath13/r@xmath2 = 0.34 model over the r@xmath13/r@xmath2 = 0.12 model with 91% confidence .
as indicated by the solar image of @xmath0 in figure [ limbs ] , a limb brightened model should be used for @xmath0 emission . the double - u model for the @xmath0 transit
fits the data better than a non - detection , even when accounting for the additional parameters in the model .
the best fit absorption radius r@xmath13/r@xmath2=0.34@xmath3 , if the absorption is entirely optically thick .
this radius favors planetary atmosphere models with mass flow beyond the planet s roche lobe .
the best time to observe transits in optically thin stellar emission lines is when a planet crosses its host s limb and not at a phase of 0 , since the double - u curve is deepest at the stellar limb .
observations at the limb have a depth of @xmath100.5 ( r@xmath13/r@xmath2)@xmath60 whereas at the center of the star they have a transit depth of @xmath100.5 ( r@xmath13/r@xmath2)@xmath61 , only half the depth of a uniform brightness transit .
the @xmath0 transit depth is comparable to the @xmath8 transit depth calculated by @xcite and larger than the @xmath54 transit depth calculated by @xcite .
as @xcite point out , a hard sphere , which we assume in equation [ disk ] , is a poor approximation to planetary atmospheric absorption . the @xmath0 absorption profile , instead of having a sharp transition from opaque to transparent , should change smoothly from optically thick to optically thin absorption in a more accurate model . in order to explain the large transit depth , this smooth model would have to have a radius extending beyond the planet s roche lobe to explain the observed transit depth .
r@xmath62/r@xmath2 varies from 0.35 to 0.48 @xcite .
the best - fit transit depth supports models with high concentrations of metallic ions in the atmosphere because the radius of @xmath0 absorption is as large as for @xmath54 absorption .
@xcite point out that if the metallicity is high , the temperature must also be elevated . the @xmath1010,000 k temperature suggested by @xcite , @xcite , and @xcite for the thermosphere may not explain the large abundances of si@xmath63 needed for the observed absorption .
the ionization energy , @xmath64 for si@xmath65 si@xmath63 = @xmath66k , suggests that the si@xmath63 may be produced in a shock between the stellar and planetary winds .
@xcite find no significant @xmath0 absorption by hd209458b .
our best fit to the stis data predicts the average transit depth to be 9@xmath67% for the same phases as their observations , but @xcite , with the cosmic origins spectrograph ( cos ) , found 0.2 @xmath68 1.4% .
these results disagree at the @xmath101.7@xmath69 level and variability in the planet s atmosphere may account for the discrepancy .
@xcite did observe some weak @xmath0 absorption features found at + 20 and + 40km / s in their spectrum , which indicates that some si@xmath63 ions remain in the planet s atmosphere or winds .
@xcite suggest that the amount of another silicon ion , si@xmath70 , may vary appreciably over short timescales because of changes in stellar wind speed , planetary mass - loss rate or temperature fluctuations .
conversely to @xmath0 , for which 2003 stis observations indicate strong absorption and 2009 cos observations indicate weak absorption , the @xmath6 absorption is seen strongly in absorption in the 2009 cos data and weakly in the 2003 stis data .
additional observations will help confirm or refute the detection of si@xmath63 in the thermosphere of hd209458b .
the cosmic origins spectrograph ( cos ) is an ideal instrument with 2 to 10 times the sensitivity of previous ultraviolet spectrographs @xcite .
cos was used by @xcite , but only when hd209458b was close to a phase of 0.0 , 0.25 , 0.5 and 0.75 and once when the planet was at the stellar limb .
additional observation at the host s limb would be optimal for @xmath0 and other optically thin emission lines . with enough signal to noise ,
the light curve may reveal asymmetries in the transit having to do with an asymmetric spatial distribution of the uv - absorbing cloud .
the advantage of the limb brightened emission lines is that they come from a smaller spatial area of the star and therefore probe the spatial distribution of the planetary atmosphere better than optically thick emission .
accurate time resolved transit data may also reveal differences in _ thermal _ properties of the leading and trailing sides of the planet as predicted by @xcite .
e. schlawin was supported by the nasa space grant fellowship .
e. agol was supported in part by the nsf under grant no .
phy05 - 51164 during a visit to the kavli institute for theoretical physics , and by nsf career grant
support for kevin covey was provided by nasa through hubble fellowship grant # hst - hf-51253.01-a awarded by the space telescope science institute , which is operated by the association of universities for research in astronomy , inc .
, for nasa , under contract nas 5 - 26555 .
lucianne walkowicz is grateful for the support of the kepler fellowship for the study of planet - bearing stars .
+ , l. , haswell , c. a. , froning , c. s. , hebb , l. , holmes , s. , kolb , u. , helling , c. , carter , a. , wheatley , p. , cameron , a. c. , loeillet , b. , pollacco , d. , street , r. , stempels , h. c. , simpson , e. , udry , s. , joshi , y. c. , west , r. g. , skillen , i. , & wilson , d. 2010 , , 714 , l222 | transit light curves for stellar continua have only one minimum and a `` u '' shape .
by contrast , transit curves for optically thin chromospheric emission lines can have a `` w '' shape because of stellar limb - brightening .
we calculate light curves for an optically thin shell of emission and fit these models to time - resolved observations of @xmath0 absorption by the planet hd209458b .
we find that the best fit @xmath0 absorption model has r@xmath1/r@xmath2 = 0.34@xmath3 , similar to the roche lobe of the planet .
while the large radius is only at the limit of statistical significance , we develop formulae applicable to transits of all optically thin chromospheric emission lines . |
einstein s theory of general relativity is the remarkably successful classical theory of the gravitational force , characterised by newton s coupling constant @xmath0 and a small cosmological constant @xmath1 .
experimentally , its validity has been confirmed over many orders of magnitude in length scales ranging from the sub - millimeter regime up to solar system size . at larger length scales , the standard model of cosmology
including dark matter and dark energy components fits the data well . at shorter length scales ,
quantum effects are expected to become important .
an order of magnitude estimate for the quantum scale of gravity the planck scale is obtained by dimensional analysis leading to the planck length @xmath2 of the order of @xmath3 , with @xmath4 the speed of light . in particle physics units
this translates into the planck mass @xmath5 while this energy scale is presently out of reach for earth - based particle accelerator experiments , fingerprints of planck - scale physics can nevertheless become accessible through cosmological data from the very early universe . from a theory perspective
, it is widely expected that a fundamental understanding of planck scale physics requires a quantum theory of gravity .
it is well known that the standard perturbative quantisation programme faces problems , and a fully satisfactory quantum theory , even outside the framework of local quantum physics , is presently not at hand . in the past 15 years , however , a significant body of work has been devoted to re - evaluate the physics of the planck scale within conventional settings .
much of this renewed interest is fueled by steven weinberg s seminal proposal , more than 30 years of age , that a quantum theory of gravitation may _ dynamically _ evade the virulent divergences encountered in standard perturbation theory @xcite .
this scenario , known as asymptotic safety , implies that gravity achieves a non - trivial ultraviolet ( uv ) fixed point under its renormalisation group flow .
if so , this would incorporate gravity alongside the set of well - understood quantum field theories whose uv behaviour is governed by a fixed point , eg . yang - mills theory . in this note ,
i discuss the renormalisation group approach to gravity .
i recall some of the issues of perturbative quantum gravity ( sec .
[ pt ] ) , introduce the renormalisation group la wilson to access the physics at the planck scale ( sec .
[ rg ] ) , review key results in four dimensions ( sec .
[ fps ] ) , evaluate applications within low - scale quantum gravity ( sec . [ extra ] ) , and conclude ( sec .
a vast body of work has been devoted to the perturbative quantisation programme of gravity .
i recall a very small selection of these in order to prepare for the subsequent discussion , and i use particle physics units @xmath6 throughout .
classical general relativity is described by the classical action @xmath7 where i have chosen a euclidean signature , @xmath8 denotes the ricci scalar , and @xmath1 the cosmological constant term .
the main point to be stressed is that the fundamental coupling of gravity @xmath9 in carries a dimension which sets a mass scale , with mass dimension @xmath10=2-d$ ] in @xmath11 dimensional space - time .
this structure distinguishes gravity in a profound manner from the other fundamentally known interactions in nature , all of which have dimensionless coupling constants from the outset .
consequently , the effective _ dimensionless _ coupling of gravity which organises its perturbative expansion is given by @xmath12 ( in four dimensions ) , where @xmath13 denotes the relevant energy scale .
while the effective coupling remains small for energies below the planck scale @xmath14 , it grows large in the planckian regime where an expansion in @xmath15 may become questionable . within the feynman diagrammatic approach
, this behaviour translates into the ( with loop order ) increasing degree of divergence of perturbative diagrams involving gravitons @xcite .
this structure is different from standard quantum field theories and relates to the classification of interactions as super - renormalisable , renormalisable , or ` dangerous ' , depending on whether their canonical mass dimension is positive , vanishing , or negative .
the degree of divergences implied by the negative mass dimension of newton s coupling is mirrored in the perturbative non - renormalisability of einstein gravity which has been established at the one - loop level @xcite in the presence of matter fields , and at the two - loop level @xcite within pure gravity . while this state of play looks discouraging from a field theory perspective , it does not rule out a quantum - field theoretical description of gravity altogether .
there are a few indicators available to support this view .
firstly , at low energies , a weak - coupling analysis of quantum gravity effects is possible within an effective theory approach @xcite , which operates an ultraviolet cutoff at the planck scale , see @xcite for a review . secondly , higher order derivative operators appear to stabilise perturbation theory @xcite . including all fourth order derivative operators it has been proven by k. stelle that gravity is renormalisable to all orders in perturbation theory @xcite .
this striking difference with einstein - hilbert gravity highlights the stabilising effect of higher derivative terms .
unfortunately , the resulting theory is not compatible with standard notions of perturbative unitarity and has therefore not been considered as a candidate for a fundamental theory of gravity .
the rle of higher order derivative operators has further been clarified in @xcite with the help of a brst analysis .
interestingly , the theory remains unitary once all higher derivative operators are retained , whereas renormalisability is at best achieved in a very weak sense due to the required infinitely many counter terms .
the renormalisation group ( rg ) comes into play when the running of couplings with energy is taken into account . as in any generic quantum field theory ,
quantum fluctuations modify the strength of couplings . if the metric field remains the fundamental carrier of the gravitational force , the fluctuations of space - time itself should modify the gravitational interactions with energy or distance . for newton s coupling
, this implies that @xmath9 becomes a running coupling @xmath16 as a function of the rg momentum scale @xmath17 , where @xmath18 denotes the wave - function renormalisation factor of the graviton .
consequently , the dimensionless coupling @xmath15 in should be replaced by the running coupling @xmath19 which evolves with the rg scale . in particular ,
the uv behaviour of standard perturbation theory is significantly improved , provided that remains finite in the high - energy limit .
this is the asymptotic safety scenario as advocated by s. weinberg @xcite , see @xcite for extensive accounts of the scenario and @xcite for brief overviews .
the intimate link between a fundamental definition of quantum field theory and renormalisation group fixed points has been stressed by k. wilson some 40 years ago @xcite . for gravity ,
the fixed point property becomes visible by considering the callan - symanzik - type rg equation for which in @xmath11 dimensions takes the form @xcite ( see also @xcite ) @xmath20 with @xmath21 the graviton anomalous dimension which in general is a function of all couplings of the theory including matter , and @xmath22 .
this simple structure arises provided the underlying effective action is local in the metric field . from the rg equation one concludes that the gravitational coupling may display two types of fixed points .
the non - interacting ( gaussian ) fixed point corresponds to @xmath23 and entails the vanishing of the anomalous dimension @xmath24 . in its vicinity
, gravity stays classical , and @xmath25 . on the other hand , a non - trivial rg fixed point with @xmath26
can be achieved implicitly , provided that the anomalous dimension reads @xmath27 the signficance of is that the graviton anomalous dimension precisely counter - balances the canonical dimension of newton s coupling @xmath9 .
this pattern is known from other gauge systems at a critical point away from their canonical space - time dimensionality @xcite , eg .
@xmath28 higgs theory in three dimensions @xcite . in consequence ,
the dimensionful , renormalised coupling scales as @xmath29 and becomes small in the ultraviolet limit where @xmath30 .
this pattern is at the root for the non - perturbative renormalisability of quantum gravity within a fixed point scenario .
much work has been devoted to check by explicit computation whether or not the gravitational couplings achieve a non - trivial uv fixed point . a versatile framework to address this question
is provided by modern ( functional ) renormalisation group methods , based on the infinitesimal integrating - out of momentum degrees of freedom from a path integral representation of the theory la wilson @xcite .
this is achieved by adding a momentum cutoff to the action , quadratic in the propagating fields . in consequence ,
the action becomes a scale - dependent effective or flowing action @xmath31 , @xmath32 + s_{k,\rm gf}+ s_{k,\rm gh } + s_{k,\rm matter}\,,\ ] ] which in the context of gravity contains a running gravitational coupling , a running cosmological constant @xmath33 , a gauge fixing term , ghost contributions , matter interactions , and the dots indicate possible higher derivative operators in the metric field . upon varying the rg scale @xmath17 , the effective action interpolates between a microscopic theory @xmath34 at the uv scale @xmath35 ( not to be confused with the running cosmological constant @xmath33 ) and the macroscopic quantum effective action @xmath36 where all fluctuations are taken into account ( @xmath37 ) .
the variation of with rg scale @xmath17 is given by an exact functional differential equation @xcite -2.5 cm @xmath38{3 cm } \includegraphics[width=4cm]{pergtalk.pdf } \end{minipage}\ ] ] -2 cm which relates the change of the scale - dependent gravitational action with the full field - dependent propagator of the theory ( full line ) and the scale - dependence @xmath39 of the momentum cutoff ( the insertion ) .
here , the trace stands for a momentum integration , and a sum over all propagating degrees of freedom @xmath40 , ghosts , matter fields ) , which minimally contains the metric field and its ghosts . the function @xmath41 ( not to be confused with the ricci scalar ) denotes the infrared momentum cutoff . as a function of ( covariant ) momenta @xmath42 ,
the momentum cutoff obeys @xmath43 for @xmath44 , @xmath45 for @xmath46 , and @xmath47 for @xmath48 ( for examples and plots of @xmath41 , see @xcite ) .
the functional flow is closely linked to other exact functional differential equations such as the callan - symanzik equation @xcite in the limit @xmath49 , and the wilson - polchinski equation by means of a legendre transformation @xcite . in a weak coupling expansion , reproduces standard perturbation theory to all loop orders @xcite .
the strength of the formalism is that it is not bounded to the weak - coupling regime , and systematic approximations such as the derivative expansion , vertex expansions , or mixtures thereof are available to access domains with strong coupling and/or strong correlation .
systematic uncertainties can be assessed @xcite , and the stability and convergence of approximations is enhanced by optimisation techniques , see @xcite . by construction ,
the flow is both ultraviolet and infrared finite .
together with the boundary condition @xmath34 it may serve as a definition for the theory . in renormalisable theories ,
the continuum limit is performed by removing the scale @xmath1 , @xmath50 , and the functional @xmath51 approaches a fixed point action at short distances ( see the proposal in @xcite for a construction in gravity ) . in
perturbatively renormalisable theories , @xmath52 is `` trivial '' and given mainly by the classical action . in
perturbatively non - renormalisable theories , the existence ( or non - existence ) of @xmath52 has to be studied on a case - by - case basis .
a viable fixed point action @xmath52 in quantum gravity should at least contain those diffeomorphism invariant operators which display relevant or marginal scaling in the vicinity of the uv fixed point .
a fixed point action qualifies as fundamental if renormalisation group trajectories @xmath53 emanating from its vicinity connect with the correct long - distance behaviour for @xmath37 and stay well - defined ( finite , no poles ) at all scales @xcite .
the operator trace in is evaluated by using flat or curved backgrounds together with heat kernel techniques or plain momentum integration .
to ensure diffeomorphism symmetry within this set - up , the background field formalism is used by adding a non - propagating background field @xmath54 @xcite .
this way , the extended effective action @xmath55 $ ] becomes gauge - invariant under the combined symmetry transformations of the physical and the background field .
a second benefit of this is that the background field can be used to construct a covariant laplacean @xmath56 ( or similar ) to define a mode cutoff at the rg momentum scale @xmath57 .
this implies that the mode cutoff @xmath41 will depend on the background fields , which is controlled by an equation similar to the flow equation itself @xcite .
the background field is then eliminated from the final equations by identifying it with the physical mean field .
this procedure dynamically readjusts the background field and implements the requirements of background independence for quantum gravity .
an alternative technique which employs a bi - metric approximation has been put forward in @xcite@xcite@xcite . for a general evaluation of the different implementations of a momentum cutoff in gauge theories ,
see @xcite , @xcite and @xcite .
in this section , i give a brief summary of fixed points found so far , and refer to @xcite for a more detailed overview of result prior to 2008 .
the search for fixed points in quantum gravity starts by restricting the running effective action @xmath31 to a finite set of operators @xmath58 with running couplings @xmath59 , @xmath60 including eg .
the ricci scalar @xmath61 and the cosmological constant @xmath62 .
the flow equations for the couplings @xmath59 are obtained from by projection onto some subspace of operators . convergence and stability of approximations is checked by increasing the number of operators retained , and further improved through optimised choices of the momentum cutoff @xcite .
0.001 ( 400,420 ) ( 20,90)@xmath63 ( -100,340)@xmath64 ( -220,190)@xmath65 ( -282,132)@xmath66 ( -380,400)@xmath67 -1 cm -2 cm the first set of flow equations which carry a non - trivial uv fixed point has been derived in @xcite for the einstein - hilbert theory in feynman gauge . as a function of the number of space - time dimensions , the flow equation reproduces the well - known fixed point in @xmath68 dimensions @xcite see fig .
[ ddep ] . more importantly , the equations display a non - trivial uv fixed point in the four - dimensional theory @xcite .
both the ricci scalar and the cosmological constant term are relevant operators at the fixed point .
their scaling is strongly correlated leading to a complex conjugate pair of universal scaling exponents .
this result has subsequently been confirmed within a more general background field gauge by means of a trace - less transverse decomposition of the metric field @xcite .
a complete analytical understanding of the einstein - hilbert approximation has been achieved in @xcite , leading to analytical flow equations and closed expressions for the ( unique ) uv fixed point and its universal eigenvalues .
key for this was the use of ( optimised ) momentum cutoffs which allow analytical access and improve convergence and stability of results @xcite .
[ vergleich ] shows the universal scaling exponents @xmath69 at the uv fixed point in four dimensions ( first column ) for various ( optimised ) choices of the momentum cutoff @xcite . also on display
is the invariant @xmath70 .
the variations in either of these are small .
the dependence on the gravitational gauge fixing parameter @xmath71 is moderate and controlled by an independent fixed point in the gauge fixing sector @xmath72 landau- de witt gauge ( @xmath73 ) @xcite . fig .
[ running ] shows the rg running of couplings along the separatrix connecting the non - trivial uv fixed point with the gaussian fixed point in the infrared . at the scale @xmath74
, the rg flow displays a cross - over from perturbative ir scaling to fixed point scaling where the anomalous dimension grows large .
the dynamical scale @xmath75 is of the order of the fundamental planck scale , and a consequence of the uv fixed point . in gravity ,
the scale @xmath75 plays a rle analogous to that of @xmath76 in quantum chromodynamics .
0.001 ( 500,900 ) ( 0,690 ) ( 0,480 ) ( 0,270 ) ( -33,30 ) ( 60,850)a)@xmath77 ( 60,640)b)@xmath78 ( 60,420)c)@xmath79 ( 60,80)d)@xmath80 ( 220,0)*@xmath11 * 0.001 ( 600,540 ) ( 400,500)@xmath81 ( 400,460)@xmath82 ( 400,380)@xmath83 ( 550,250)@xmath84 ( 120,380 ) ir ( 480,380 ) uv running couplings along the separatrix in four dimensions ; from @xcite.,title="fig : " ] -4 cm in a formidable tour - de - force computation , the space of operators has been extended to include @xmath85 interactions @xcite .
the perturbatively marginal @xmath86 coupling turned out to become relevant at a uv fixed point .
elsewise , it added only minor corrections to the fixed point in the einstein - hilbert theory . with @xcite ,
these computations became feasible analytically and have been implemented for operators @xmath87 which contain a general function of the ricci scalar in @xcite and @xcite .
a non - trivial uv fixed point has been found by taylor - expanding @xmath88 to high orders in the ricci scalar with results known up to eighth order in the eigenvalues @xcite and up to tenth order in the fixed point @xcite . additionally , the fixed is remarkably stable .
the universal eigenvalues converge rapidly , confirming all previous results .
most importantly , it turned out that operators @xmath89 from @xmath90 onwards become _ irrelevant _ at the uv fixed point @xcite , a first indicator that an asymptotically safe uv fixed point of the fundamental theory may in fact have a finite number of relevant operators .
this is most welcome as otherwise a fixed point with infinitely many relevant couplings is likely to spoil the predictive power .
all fourth order derivative operators have been included both at the one - loop order and beyond .
the re - evaluation of perturbative one - loop results by means of a functional flow emphasized the significance of eg .
quadratic divergences which normally are suppressed within dimensional regularisation @xcite .
two types of uv fixed points are found .
the first one is non - trivial in all couplings @xcite with three relevant directions at the fixed point .
the second one is perturbative in the weyl coupling @xcite but non - trivial in newton s coupling , with two relevant and two marginally relevant directions . in either case newton s
coupling takes a non - trivial fixed point in the uv .
these result hence indicate that higher derivative operators are compatible with gravity developing a non - trivial fixed point . once this is achieved , the difference between the einstein - hilbert approximation and the fourth order approximation is qualitatively small . in the above ,
the rg running of the ghost sector is expected to be subleading .
this has recently been confirmed in @xcite and in @xcite , and for the scalar curvature - ghost coupling in @xcite .
an important extension deals with the inclusion of matter couplings . in eg .
yang - mills theory , it is well - known that the property of asymptotic freedom can be spoiled by too many species of fermions . in the same vein it has to be checked whether the uv fixed point of gravity remains stable under the inclusion matter , and vice versa . by
now , this question has been analysed for minimally coupled scalar matter @xcite , non - minimally coupled scalar fields @xcite , and generic free matter coupled to gravity @xcite .
these findings are consistent with earlier results using different techniques in the limit of many matter fields @xcite@xcite .
the conclusion is that the gravitational fixed point generically persists under the inclusion of matter ( including the standard model and its main extensions ) unless a significant imbalance between bosonic and fermionic matter fields is chosen . also , non -
minimally coupled scalar matter leads to slight deformation of the gravitational fixed point with matter achieving a weakly coupled regime , the so - called gaussian matter fixed point .
fermions have been included recently in @xcite , as well as gravitational yukawa systems @xcite . in the latter case ,
the coupled system displays a non - trivial uv fixed point both in the yukawa as well as in the gravitational sector .
this new fixed point may become of great interest as a stabilizer for the standard model higgs .
on the one - loop level , the interplay between a standard model higgs and asymptotically safe gravity has been addressed in @xcite , detailing conditions under which no new physics is required to bridge the energy range from the electroweak scale all the way up to . in a similar spirit
, it has been argued that the consistency of higgs inflation models is enhanced provided that gravity becomes asymptotically safe @xcite .
the impact of gravitational fluctuations on gauge theories has seen re - newed interest initiated by a one - loop study within effective theory @xcite .
asymptotic freedom is sustained by gravity leading to a vanishing @xcite or non - vanishing one - loop correction @xcite , depending on the regularisation . in the present framework ,
, the gravity - induced corrections have been studied in @xcite and in @xcite . in either case
the sign of the graviton contributions is fixed and asymptotic freedom of yang - mills theory persists in the presence of gravitational fluctuations , also including a cosmological constant .
furthermore , @xcite evaluates generic regularisations and clarifies the non - universal nature of the one - loop coefficient , also covering generic field - theory - based uv scenarios for gravity including asymptotically safe gravity and gravitational shielding .
an interesting consequence of a non - trivial gravitational contribution to the running of an abelian charge is the appearance of a combined uv fixed point in the @xmath28-gravity system providing a mechanism to calculate the fine structure constant @xcite . in three dimensions ,
a fixed point study has been reported within topological massive gravity which includes a chern - simons ( cs ) term in addition to newton s coupling and the cosmological constant @xcite . on the one - loop level
, the cs term has a vanishing beta function and the gravitational sector displays both a gaussian and a nontrivial fixed point for positive cs coupling which is qualitatively in accord with the fixed point in fig .
[ ddep ] ( no cs term ) .
an interesting structural link has been noted between the gravitational fixed point @xcite ( ricci scalar , no cosmological constant ) and a uv fixed point in non - linear sigma models @xcite . in either system
the beta - function of the relevant coupling displays a uv fixed point whose coordinates are non - universal .
however , the systems display the same universal scaling , with identical exponent @xmath91 for all @xmath92 dimensions .
this coincidence highlights that non - linear sigma models have a non - trivial uv dynamics on their own , possibly quite similar to gravity itself , which is worth of being explored further @xcite .
fixed points in more complex ( gauged and non - gauged ) non - linear sigma models have recently been obtained in @xcite .
it would be very helpful to understand which mechanism implies the gravitational fixed point by eg . restricting the dynamical content .
in fact , a two killing vector reduction of four - dimensional gravity ( with scalar matter and photons ) has been shown to be asymptotically safe @xcite .
this topic has also been studied by reducing gravity to the dynamics of its conformal sector @xcite including higher derivative terms and effects from the conformal anomaly @xcite .
although a uv fixed point is visible in most approximations , its presence depends more strongly on eg .
matter fields , meaning that more work is required to settle this question for pure gravity .
implications of a gravitational fixed point have been studied in the context of black holes , astrophysics and infrared gravity , and cosmology .
the rg induced running @xmath93 of newton s coupling has interesting implications for the existence of rg improved black hole space - times @xcite as well as for their dynamics @xcite . if gravity weakens according to its asymptotically safe rg flow , results point towards the existence of a smallest black hole with critical mass @xmath94 of the order of @xmath95 .
this pattern persists for schwarzschild black holes in higher dimensions @xcite , rotating black holes @xcite , and black holes within higher derivative gravity @xcite , showing that the existence of smallest black holes is a generic and stable prediction of this framework .
low energy implications of the rg running of couplings have equally been studied , including applications to galaxy rotation curves @xcite .
recent studies have re - evaluated gravitational collapse @xcite exploring the one - loop running couplings , and the rg induced non - local low energy corrections to the gravitational effective action @xcite .
it has also been conjectured that gravity may possess an infrared fixed point . in this case , newton s coupling grows large at large distances and may contribute to an accelerated expansion of the universe , a scenario which can be tested against cosmological data such as type ia supernov @xcite .
the intriguing idea that an asymptotically safe uv fixed point may control the beginning of the universe has been implemented in einstein gravity with an ideal fluid @xcite , and in the context of @xmath96-theories of gravity @xcite .
the rg scale parameter is linked with cosmological time and therefore leads to rg improved cosmological equations @xcite , which may even generate entropy @xcite .
the framework put forward in @xcite includes scalar fields and exploits the bianchy identity , leading to general conditions of existence for cosmological fixed points ( with cosmic time ) whose solutions include eg . inflating uv safe fixed points of gravity coupled to a scalar field .
the proposal that lorentz symmetry can be broken by gravity on a fundamental level has been put forward in @xcite where space and time scale differently in the uv with a relative dynamical exponent @xmath97 . provided @xmath98 ( in four dimensions )
gravity is power - counting renormalisable in perturbation theory and lorentz symmetry should emerge as a low energy phenomenon @xcite .
interestingly , horava - lifshitz gravity with @xmath99 also requires an asymptotically safe fixed point for gravity with an anomalous dimension @xmath100 compensating for power - counting non - renormalisability .
the rg scaling in the vicinity of an asymptotically safe fixed point implies scale invariance . on the other hand
, it is often assumed that a quantum theory of gravity should , in one way or another , induce a minimal length .
different aspects of this question and interrelations with other approaches to quantum gravity have been addressed in @xcite , and more recently in @xcite and @xcite .
the findings indicate that the fixed point _ implicitly _ induces a notion of minimal length , related to the rg trajectory and fundamental scale where gravity crosses over from classical to fixed point scaling .
it has also been argued that rg corrections to gravity can modify the dispersion relation of massive particles leading to indirect bounds on the one - loop coefficients of the gravitational beta - functions @xcite .
finally , it is worth noting that standard quantum field theories ( without gravity ) may become asymptotically safe in their own right .
this has first been exemplified for perturbatively non - renormalisable gross - neveu models in three dimensions with functional rg methods @xcite and in the limit of many fermion flavours @xcite . in four dimensions ,
this possibility has been explored for variants of the standard model higgs @xcite , for yukawa - type interactions @xcite , and strongly coupled gauge theories @xcite .
it has also been conjectured that the electroweak theory without a dynamical higgs field could become asymptotically safe @xcite , based on structural similarities with quantum general relativity . in summary , there is an increasing amount of evidence for the existence of non - trivial uv fixed points in four - dimensional quantum field theories including gravity and matter .
( 1000,20 ) 0.001 and a dynamical planck scale @xmath75 with @xmath101 .
the fixed point behaviour in the deep ultraviolet enforces a softening of gravity .
a ) at the scale @xmath102 the coupling displays a cross - over from fixed point scaling to classical scaling in the higher - dimensional theory .
b ) at the compactification scale @xmath103 , the @xmath104 compactified spatial dimensions are no longer available for gravity to propagate in , and the running coupling displays a cross - over from @xmath105-dimensional to @xmath106-dimensional scaling.,title="fig : " ] ( 1000,5 ) ( 300,340)@xmath107 ( 570,280)uv fixed point ( 350,80)ir ( 510,260)a ) ( 430,120)b ) ( 710,20)@xmath108 ( 530,0)@xmath109 ( 430,0)@xmath110
next i turn to fixed points of quantum gravity in @xmath111 dimensions where @xmath104 denotes the number of extra dimensions @xcite , and potential signatures thereof in models with a low quantum gravity scale @xcite .
there are several motivations for this .
firstly , particle physics models where gravity lives in a higher - dimensional space - time have raised enormous interest by allowing the fundamental planck scale to be as low as the electroweak scale @xcite , @xmath112 this way , the notorious hierarchy problem of the standard model ( sm ) is circumnavigated .
if realized in nature , this opens the exciting possibility that planck scale physics becomes accessible to experiment via eg .
high - energetic particle collisions at the lhc .
the main point of these models is that the four - dimensional planck scale is no longer fundamental but a derived quantity , related as @xmath113 to the planck scale of the higher - dimensional theory and the size of the compact extra dimensions @xmath114 . secondly , the critical dimension of gravity the dimension where the gravitational coupling has vanishing canonical mass dimension is two .
hence , for any dimension above the critical one , the canonical dimension of newton s coupling is negative . from a renormalisation group point of view , this means that four dimensions appear not to be special .
continuity in the dimension suggests that an ultraviolet fixed point , if it exists in four dimensions , should persist towards higher dimensions .
furthermore , gravity has @xmath115 propagating degrees of freedom which rapidly increases with increasing dimensionality .
it is important to understand whether these additional degrees of freedom spoil or support the uv fixed point detected in the four - dimensional theory . the local structure of quantum fluctuations , and hence local renormalisation group properties of a quantum theory of gravity , are qualitatively similar for all dimensions above the critical one , modulo topological effects for specific dimensions .
therefore one should expect to find similarities in the ultraviolet behaviour of gravity in four and higher dimensions .
0.001 ( 1000,280 ) this expectation has been confirmed in @xcite for @xmath11-dimensional einstein - hilbert gravity .
the stability of the result has subsequently been tested through an extended fixed points search in higher - dimensional gravity for general cutoffs @xmath41 @xcite , also probing the stability of the fixed point against variations of the gauge fixing @xcite , and the results are displayed in fig .
[ vergleich ] .
the cutoff variations are very moderate , and smaller than the variation with gauge fixing parameter . the structural stability of the fixed point
also strengthens the findings in the four - dimensional case .
thus , the following picture emerges : in the setting with compact extra dimensions the running dimensionless coupling @xmath81 displays a cross - over from fixed point scaling to @xmath11-dimensional classical scaling at the cross - over scale @xmath75 of the order of @xmath116 , see fig .
[ runningg]a ) . by construction , this has to happen at energies way above @xmath117 . at lower energies , close to the compactification scale @xmath118 ,
the @xmath104 compactified spatial dimensions are no longer available for gravity to propagate in , and the running coupling displays a cross - over from @xmath105-dimensional to @xmath106-dimensional scaling see fig .
[ runningg]b ) .
high energetic particle colliders such as the lhc are sensitive to the electroweak energy scale , and hence to the regime of fig .
[ runningg]a ) , provided that the fundamental planck scale is of the order of .
a generic prediction of large extra dimensions is a tower of massive kaluza - klein gravitons @xcite . in particular
, the exchange of virtual gravitons in drell - yan processes will lead to deviations in standard model reference processes such as @xmath119 @xcite .
gravitational drell yan production is mediated through tree - level graviton exchange and via graviton loops .
within effective theory , the corresponding effective operators are ultraviolet divergent ( in two or more extra dimensions ) and require an uv cutoff . within fixed point gravity ,
the graviton is dressed by its anomalous dimension leading to finite and cutoff independent results for cross sections @xcite , see @xcite for a comparison of different uv completions .
[ fig : discovery ] displays the @xmath120 discovery reach to detect the fundamental scale of gravity assuming @xmath121 and an integrated luminosity of @xmath122 at the lhc .
note that the curves are given as functions of an energy cutoff @xmath1 .
in effective theory , these curves continue to grow with @xmath1 , meaning that the cutoff can not be removed . here , the leveling - off is a dynamical consequence of the underlying fixed point , and the amplitude is independent of any cutoff .
[ fig : sm ] shows the differential cross - sections for di - muon production within asymptotically safe gravity , with cross - over scale @xmath123 @xcite .
for all dimensions considered , the enhanced di - lepton production rate sticks out over standard model backgrounds , leading to the conclusion that drell yan production is very sensitive to the quantum gravity scale . within asymptotically safe gravity for @xmath124 tev and @xmath125 ( magenta ) , @xmath126 ( blue ) and @xmath127 ( red ) extra dimensions , in comparison with standard model background ( black dashed line ) ; from @xcite.,title="fig : " ] ( -245,45 ) ( -10,10)@xmath128 ( -145,60)standard model @xmath129 ( -91,157)(2,0)10 ( -75,155)@xmath125 ( -91,147)(2,0)10 ( -75,145)@xmath126 ( -91,137)(2,0)10 ( -75,135)@xmath127 ( -110,115)@xmath130 fixed point further collider signatures of fixed point gravity in warped and large extra dimensions have been addressed in @xcite within a form - factor approximation @xcite . modifications to the semi - classical production of planck - size black holes at the lhc induced by the fixed point have been analysed in @xcite .
the main new effect is an additional suppression of the production cross section for mini - black holes .
some properties of planck - size black holes have also been evaluated within a one - loop approximation , provided the gravitational rg flow is driven by many species of particles @xcite .
graviton loop corrections to electroweak precision observables within asymptotically safe gravity with extra dimensions have recently been obtained in @xcite . in summary ,
experimental signatures for the quantisation of gravity within the asymptotic safety scenario are in reach for particle colliders , provided the fundamental scale of gravity is as low as .
renormalisation group methods have become a key tool in the attempt to understand gravity at shortest , and possibly at largest , distances .
it is conceivable that planckian energies can be assessed with the help of running gravitational couplings such as , replacing from perturbation theory .
if gravity becomes asymptotically safe , its ultraviolet fixed point acts as an anchor for the underlying quantum fluctuations .
the increasing amount of evidence for a gravitational fixed point with and without matter and its significance for particle physics and cosmology are very promising and certainly warrant farther reaching investigations .
s. weinberg , ultraviolet divergences in quantum theories of gravitation . in _ general relativity : an einstein centenary survey _ ,
s.w . hawking and w. israel , cambridge university press ( 1979 ) ,
p. 790 .
j. f. donoghue , leading quantum correction to the newtonian potential .
lett . * 72 * ( 1994 ) 2996 [ gr - qc/9310024 ] .
c. p. burgess , quantum gravity in everyday life : general relativity as an effective field theory .
living rev .
* 7 * ( 2004 ) 5 [ gr - qc/0311082 ] .
m. niedermaier , the asymptotic safety scenario in quantum gravity : an introduction . class .
* 24 * ( 2007 ) r171 [ gr - qc/0610018 ] .
k. s. stelle , renormalization of higher derivative quantum gravity .
phys . rev . * d16 * ( 1977 ) 953 - 969 .
d. f. litim , on fixed points of quantum gravity .
aip conf .
proc . * 841 * ( 2006 ) 322 [ hep - th/0606044 ] .
r. percacci , asymptotic safety . in oriti , d. ( ed . ) : approaches to quantum gravity 111 - 128 .
[ arxiv:0709.3851 [ hep - th ] ] . d. f. litim .
fixed points of quantum gravity and the renormalisation group .
pos * qg - ph * 024 ( 2008 ) , 0810.3675 [ hep - th ]
. k. g. wilson , renormalization group and critical phenomena .
1 . renormalization group and the kadanoff scaling picture .
* b4 * ( 1971 ) 3174 - 3183 .
j. berges , n. tetradis and c. wetterich , non - perturbative renormalisation flow in quantum field theory and statistical physics .
* 363 * ( 2002 ) 223 [ hep - ph/0005122 ] .
t. r. morris , the exact renormalization group and approximate solutions .
j. mod .
* a9 * ( 1994 ) 2411 - 2450 .
[ hep - ph/9308265 ] .
m. reuter , nonperturbative evolution equation for quantum gravity .
d * 57 * ( 1998 ) 971 [ hep - th/9605030 ] .
d. f. litim , optimisation of the exact renormalisation group .
phys . lett . *
b486 * ( 2000 ) 92 [ hep - th/0005245 ] . k. symanzik , small distance behavior in field theory and power counting .
phys . * 18 * ( 1970 ) 227 .
d. f. litim and j. m. pawlowski , perturbation theory and renormalisation group equations .
d * 65 * ( 2002 ) 081701 [ hep - th/0111191 ] .
d. f. litim and j. m. pawlowski , completeness and consistency of renormalisation group flows .
d * 66 * ( 2002 ) 025030 [ hep - th/0202188 ] .
d. f. litim , d. zappal , ising exponents from the functional renormalisation group .
[ arxiv:1009.1948 [ hep - th ] ] .
d. f. litim , critical exponents from optimised renormalisation group flows .
. phys .
b * 631 * ( 2002 ) 128 [ hep - th/0203006 ] .
j. m. pawlowski , aspects of the functional renormalisation group .
annals phys .
* 322 * ( 2007 ) 2831 [ hep - th/0512261 ] .
e. manrique , m. reuter , bare action and regularized functional integral of asymptotically safe quantum gravity .
* d79 * ( 2009 ) 025008 .
[ arxiv:0811.3888 [ hep - th ] ] .
f. freire , d. f. litim and j. m. pawlowski , gauge invariance and background field formalism in the exact renormalisation group .
b * 495 * , 256 ( 2000 ) [ hep - th/0009110 ] .
d. f. litim and j. m. pawlowski , wilsonian flows and background fields .
b * 546 * ( 2002 ) 279 [ hep - th/0208216 ] .
s. folkerts , d. f. litim , j. m. pawlowski , asymptotic freedom of yang - mills theory with gravity .
[ 1101.5552 [ hep - th ] ] .
e. manrique and m. reuter , bimetric truncations for quantum einstein gravity and asymptotic safety .
annals phys .
* 325 * ( 2010 ) 785 [ arxiv:0907.2617 [ gr - qc ] ] .
e. manrique , m. reuter and f. saueressig , matter induced bimetric actions for gravity .
arxiv:1003.5129 [ hep - th ] .
e. manrique , m. reuter and f. saueressig , bimetric renormalization group flows in quantum einstein gravity .
arxiv:1006.0099 [ hep - th ] .
r. gastmans , r. kallosh and c. truffin , quantum gravity near two - dimensions . nucl .
b * 133 * ( 1978 ) 417 s. m. christensen and m. j. duff , quantum gravity in two + epsilon dimensions .
b * 79 * ( 1978 ) 213 .
o. lauscher and m. reuter , ultraviolet fixed point and generalized flow equation of quantum gravity .
d * 65 * ( 2002 ) 025013 [ hep - th/0108040 ] .
d. f. litim , fixed points of quantum gravity .
lett . * 92 * ( 2004 ) 201301 [ hep - th/0312114 ] .
fischer and d. f. litim , fixed points of quantum gravity in extra dimensions . phys .
b * 638 * ( 2006 ) 497 [ hep - th/0602203 ] .
d. f. litim , j. m. pawlowski , flow equations for yang - mills theories in general axial gauges .
* b435 * ( 1998 ) 181 - 188 .
[ hep - th/9802064 ] . .
o. lauscher and m. reuter , flow equation of quantum einstein gravity in a higher - derivative truncation .
d * 66 * ( 2002 ) 025026 [ hep - th/0205062 ] .
o. lauscher and m. reuter , is quantum einstein gravity nonperturbatively renormalizable ?
* 19 * ( 2002 ) 483 [ hep - th/0110021 ] .
a. codello , r. percacci and c. rahmede , ultraviolet properties of f(r)-gravity .
j. mod .
a * 23 * ( 2008 ) 143 [ 0705.1769 [ hep - th ] ] .
a. codello , r. percacci , c. rahmede , investigating the ultraviolet properties of gravity with a wilsonian renormalization group equation .
annals phys .
* 324 * ( 2009 ) 414 - 469 .
[ arxiv:0805.2909 [ hep - th ] ] .
p. f. machado and f. saueressig , on the renormalization group flow of f(r)-gravity .
d * 77 * ( 2008 ) 124045 [ 0712.0445 [ hep - th ] ] .
a. codello and r. percacci , fixed points of higher derivative gravity .
* 97 * ( 2006 ) 221301 [ hep - th/0607128 ] .
m. r. niedermaier , gravitational fixed points from perturbation theory .
* 103 * ( 2009 ) 101303 .
r. percacci and d. perini , asymptotic safety of gravity coupled to matter .
d * 68 * ( 2003 ) 044018 [ hep - th/0304222 ] .
e. tomboulis , 1/n expansion and renormalization in quantum gravity .
* b70 * ( 1977 ) 361 .
d. j. toms , quantum gravity and charge renormalization .
d * 76 * , 045015 ( 2007 ) [ arxiv:0708.2990 [ hep - th ] ] .
d. ebert , j. plefka and a. rodigast , absence of gravitational contributions to the running yang - mills coupling .
b * 660 * , 579 ( 2008 ) [ arxiv:0710.1002 [ hep - th ] ] .
m. reuter , h. weyer , conformal sector of quantum einstein gravity in the local potential approximation : non - gaussian fixed point and a phase of unbroken diffeomorphism invariance .
* d80 * ( 2009 ) 025001 .
[ arxiv:0804.1475 [ hep - th ] ] .
d. f. litim and t. plehn , signatures of gravitational fixed points at the lhc .
* 100 * ( 2008 ) 131301 [ 0707.3983 [ hep - ph ] ] .
d. f. litim and t. plehn , virtual gravitons at the lhc .
0710.3096 [ hep - ph ]
. j. hewett and t. rizzo , collider signals of gravitational fixed points .
jhep * 0712 * ( 2007 ) 009 [ 0707.3182 [ hep - ph ] ] .
e. gerwick , d. f. litim , t. plehn , asymptotic safety and kaluza - klein gravitons at the lhc .
[ arxiv:1101.5548 [ hep - ph ] ] .
n. arkani - hamed , s. dimopoulos and g. r. dvali , the hierarchy problem and new dimensions at a millimeter . phys .
b * 429 * , 263 ( 1998 ) [ arxiv : hep - ph/9803315 ] .
i. antoniadis , n. arkani - hamed , s. dimopoulos and g. r. dvali , new dimensions at a millimeter to a fermi and superstrings at a tev .
b * 436 * , 257 ( 1998 ) [ arxiv : hep - ph/9804398 ] .
l. randall and r. sundrum , a large mass hierarchy from a small extra dimension .
* 83 * , 3370 ( 1999 ) [ arxiv : hep - ph/9905221 ]
. g. f. giudice , r. rattazzi , j. d. wells , quantum gravity and extra dimensions at high - energy colliders .
* b544 * ( 1999 ) 3 - 38 .
[ hep - ph/9811291 ] .
t. han , j. d. lykken and r. j. zhang , on kaluza - klein states from large extra dimensions .
d * 59 * ( 1999 ) 105006 , [ hep - ph/9811350 ] .
j. l. hewett , indirect collider signals for extra dimensions .
* 82 * ( 1999 ) 4765 , [ hep - ph/9811356 ] .
r. k. ellis , w. j. stirling and b. r. webber , qcd and collider physics .
monogr . part .
cosmol .
* 8 * , ( 1996 ) 1 .
e. gerwick and t. plehn .
extra dimensions and their ultraviolet completion .
[ arxiv:0912.2653 [ hep - ph ] ] . | i discuss the renormalisation group approach to gravity , its link to s. weinberg s asymptotic safety scenario , and give an overview of results with applications to particle physics and cosmology . |
in recent years much progress has been made in the field of freeform surface design without the assumption of symmetries @xcite .
the goal of these design methods is the solution of the so called inverse problem of nonimaging optics .
this means that for given _ arbitrary _ source and target intensities @xmath0 and @xmath1 one or more freeform surfaces have to be calculated , which map the intensities onto each other . especially the design of _ continuous _ freeform surfaces , on which we will concentrate on in the following , is a challenging problem and of great interest for practical applications .
+ the first successful method at calculating a continuous freeform surface utilizing a complex target intensity was developed by ries and muschaweck @xcite , but unfortunately the numerical method was not published @xcite .
their approach is able to handle the far field design problem for single freeform surfaces illuminated by a point source @xcite .
+ nowadays , many other methods have been developed by different research groups . a quite popular approach
are the so - called monge - ampre methods @xcite .
they are based on the modelling of the design problem by a nonlinear partial differential equation of monge - ampre type and solving it with sophisticated numerical techniques .
these methods are able to handle the design problem in the far field for intensity control of point sources @xcite and collimated beams @xcite as well as intensity and phase control with double freeform surfaces @xcite . +
another popular approach for the single freeform surface design with point sources is the supporting ellipsoids method developed by oliker @xcite . with this method
a freeform mirror is constructed by putting a point source in the focal point of a unification of ellipoids , whereby every ellipsoid has a different position of the second focal point on the target plane to build the required intensity pattern .
the challenge of this method is the calculation of a smooth freeform surface by the unification of the ellipsoids .
therefore it was further developed by other research groups @xcite and generalized to calculate freeform lenses @xcite .
it can handle the far field as well as the near field design problem .
+ also quite often used are ray - mapping techniques @xcite , which are frequently based on the calculation of a ray mapping between the source and the target intensity and a subsequent construction of the freeform surface @xcite .
the aim and challenging part of these methods is to find an _ integrable _ ray mapping , which allows the calculation of a _ continuous _ surface .
+ the approach we will concentrate on is a subgroup of the ray - mapping techniques , which are called optimal mass transport ( omt ) methods @xcite .
these gained some interest in recent years and are partly based on the mathematical concept of optimal mass transport as explained in the following paragraph .
they can handle the design problem of a single and double freeform surfaces for intensity control @xcite as well as double freeform surfaces for intensity and phase control @xcite .
the connection between mass transport and freeform design was also discussed in @xcite . +
this approach for the freeform surface design consists of two separate steps . in the first step a ray mapping between
the given input and output intensities is calculated via omt . in the second step
the freeform surface is constructed with the help of the law of refraction / reflection and the well - known integrability condition , which ensures the continuity of the surface .
+ the difference between the omt methods mentioned above is the second step . in first approach by buerle _
the freeform is constructed by an optimization procedure @xcite , while the second approach by feng _
et al . _ uses a simultaneous point - by - point construction method to design a double freeform surface @xcite .
+ since these attempts seem to be quite successful but do not give theoretical insights about the integrability of the omt map , we want to clarify this point in our work for a single freeform illuminated by a collimated beam .
this will be done by deriving a condition for an integrable map and showing that it can be fulfilled ( approximately ) by the omt map .
based on our findings , we present an efficent and easy - to - implement numerical freeform surface construction technique differing from the previously published omt methods @xcite . to do so this paper is structured as follows .
+ in section [ sec : design method ] after a short introduction to the omt and a presentation of its basic properties , we will derive from the law of reflection / refraction and the integrability condition a _ general _ condition for an integrable ray mapping and its corresponding surface for collimated input beams .
it will be shown that in a small - angle approximation this condition can be fulfilled by using an omt mapping and therefore the freeform surface design process indeed decouples into the two steps described above .
thereby it will be shown that the freeform surface can be constructed from a linear advection equation with appropriate boundary conditions . in section [ sec : numerical algorithm ] we then argue that the advection equation for the freeform construction can be solved by simple integrations , which is different to the omt freeform design methods mentioned above and implies the approximate integrability over a wide range of freeform - target plane distances . the efficiency of this approach is then demonstrated in section [ sec : examples ] by applying it to two challenging design examples , followed by a discussion of our results in section [ sec : conclusion ] .
the problem statement of omt , also called the monge - kantorovich problem , is as follows : two positive density functions @xmath2 and @xmath3 with @xmath4 have to be mapped onto each other according to the jacobi equation @xmath5 with a smooth , bijective mapping @xmath6 .
if @xmath7 is defined as the set of mappings fulfilling equation ( [ eq:2 ] ) , we are searching for a mapping minimizing the transport cost according to the kantorovich - wasserstein distance @xmath8 whereby @xmath9 denotes the mapping for which the integral has its minimal value .
this mapping , which is defined by ( [ eq:1 ] ) , ( [ eq:2 ] ) and ( [ eq:3 ] ) , has the useful property that it is unique @xcite and it is characterized by a vanishing curl @xcite .
the latter property will be important for our findings in the next subsection .
+ in the special case of freeform surface design considered here , the densities @xmath2 and @xmath3 correspond to the source and target intensities with the units @xmath10 ( see fig . [
fig : geometrie ] ) . therefore equation ( [ eq:1 ] ) describes a global and equation ( [ eq:2 ] ) a local energy conservation .
+ for the numerical examples in section [ sec : examples ] , we have implemented the omt method developed by sulman _
it provides a good compromise between an easy implementation and an efficient mapping calculation and is thus sufficient for our test purposes .
+ the result of this design process step is therefore the mapping @xmath11 , but we have to keep in mind , that it is not obvious at this point , whether the omt mapping is integrable and if it is , for which lens - target distances ( see fig . [
fig : geometrie ] ) this is the case .
this point will be clarified in the next subsection . in the following ,
we want to derive a differential equation for the direct calculation of a freeform surface for a given _ general _ ray mapping @xmath11 .
our derivations will lead to a condition an _ integrable _ mapping and its corresponding freeform surface have to fulfill .
as we will show , this condition can be fulfilled approximately over a wide range of lens - target distances by the omt mapping defined in the previous subsection .
+ to do so , two basic equations are considered . on the one hand , for an incoming beam described by the ray direction vector field @xmath12 and the refracted vector field @xmath13 , the law of refraction @xmath14 with the refractive indices @xmath15 of the lens and @xmath16 of the surrounding medium , has to be fulfilled . on the other hand , we want to ensure the continuity of the surface @xmath17 by the well - known integrability condition @xcite @xmath18 since the collimated beam @xmath12 as well as @xmath13 can be expressed in terms of the unknown freeform surface @xmath17 and the given ray mapping @xmath11 ( see fig .
[ fig : geometrie ] ) : @xmath19 the equations ( [ eq:4 ] ) and ( [ eq:5 ] ) represent a differential equation for @xmath17 . plugging ( [ eq:4 ] ) into equation ( [ eq:5 ] ) the integrability condition can be written in the form ( see appendix a ) @xmath20\}_z } { \mathbf{n}\cdot\mathbf{s}_2 } + \mathbf{s}_2(\nabla \times \mathbf{s}_3),\ ] ] which holds for a _ general _ ray mapping @xmath11 .
equation ( [ eq:7 ] ) is organized in a way that only the left - hand side ( lhs ) depends on the derivatives of @xmath17 .
equation ( [ eq:7 ] ) takes a more familiar form by inserting the vector fields ( [ eq:6 ] ) , which leads to @xmath21}{\mathbf{n}\cdot \mathbf{s}_2}-(z_t- z(x , y ) ) \nabla \mathbf{v},\ ] ] with the velocity field @xmath22 , the identity vector @xmath23 and @xmath24 .
this equation is a semilinear two dimensional advection equation , whereby the unknown surface @xmath17 corresponds to conserved transport quantity and the right - hand side ( rhs ) to a source term .
+ in principle one could try to solve equation ( [ eq:8 ] ) after applying suitable boundary conditions , but as it will be demonstrated , it is not appropriate for finding a continuous freeform surface .
this can be seen easily by considering the condition that the normalized vector field ( [ eq:4 ] ) has to be equal to the gradient of the surface @xmath17 : @xmath25 plugging this relation into the lhs of ( [ eq:8 ] ) , we get @xmath26 , which can only be fulfilled if the rhs vanishs : @xmath27}{\mathbf{n}\cdot \mathbf{s}_2}-(z_t- z(x , y ) ) \nabla \mathbf{v } \stackrel{!}{\equiv}0.\ ] ] the importance of this condition is due to the fact that is has to hold for _ every _ integrable ray mapping . since @xmath28 and therefore @xmath29 it reflects the nature of the law of refraction , that according to the definitions ( [ eq:6 ] ) and ( [ eq:9 ] ) the vectors @xmath30 , @xmath31 and @xmath32 have to lie in the same plane . + we now know , that the source term of ( [ eq:8 ] ) has to vanish , but we are still left with the question , if we can find a way to fullfil condition ( [ eq:10 ] ) .
the main task is obviously to find a ray mapping for which relation ( [ eq:10 ] ) holds , which is nontrivial , since it couples the mapping with the unknown function @xmath17 . +
but if we use the omt mapping , it follows from its vanishing curl that @xmath33 .
if we use in addition to that , the small - angle approximation @xmath34\ ] ] between the surface normal @xmath32 and the outgoing ray @xmath35 , we see that the condition ( [ eq:10 ] ) can be fulfilled approximately by using an omt map .
because of the fact that @xmath36 is locally proportional to @xmath37 in contrast to the rhs , ( [ eq:11 ] ) can be interpreted as a far field approximation .
this also implies , that the integrability of omt map is only asymptotically exact .
+ hence for an omt mapping and the small - angle approximation , we get our final equation @xmath38 which has to be solved to get the required freeform surface @xmath17 .
+ if we want to solve a linear advection equation like ( [ eq:12 ] ) , we have to know the function @xmath17 on the inflow part of the boundary , where the velocity field points into the integration area @xmath39 @xcite .
because of energy conservation , this area is defined as @xmath40 and therefore its inflow part by @xmath41 with the outward boundary normal @xmath42
. together with this boundary condition equation ( [ eq:12 ] ) has at most one solution @xcite .
+ in our case the boundary conditions can be deduced in the following way .
first , we have to realize , that for an incoming collimated beam the boundary of the freeform surface can only determine the tangential deflection of a ray refracted at the boundary .
the normal deflection is determined by the inner parts of the surface @xmath17 .
therefore it makes sense to parameterize the boundary @xmath43 by a parameter @xmath44 and define the local coordinate system at each point of the boundary by the vectors(see fig .
[ fig : boundary ] ) @xmath45 since @xmath46 only determines the tangential deflection , it is sufficient to consider the law of refraction ( [ eq:4 ] ) in the tangential plane spanned by @xmath47 and @xmath48 . hence , we can interpret the boundary value calculation as a two dimensional problem which allows us to derive a differential equation for the boundary values @xmath49 by the two dimensional equivalent of equation ( [ eq:9 ] ) projected on the @xmath50 plane @xmath51 where path length @xmath52 was introduced for dimensional reasons . from this
we get : @xmath53 which reduces in the far field to @xmath54 whereby the position of the surface in space compared to the target plane is fixed by integration constant . since ( [ eq:12 ] ) itself can be interpreted as a far field approximation , as explained above , equation ( [ eq:16 ] )
seems more suitable for our purposes .
it provides us with a simple way of calculating the boundary values .
the only degree of freedom left is the integration constant .
equation ( [ eq:16 ] ) will build the basis of the numerical algorithm for solving equation ( [ eq:12 ] ) presented in the next section .
+ at the end of this section , we want to note that ( [ eq:12 ] ) and ( [ eq:16 ] ) can be derived analogously for freeform mirrors by replacing ( [ eq:4 ] ) by the law of reflection and keeping in mind that @xmath55
. is parameterized by @xmath44 . at each point of the boundary the local coordinate system
is spanned by the tangential vector @xmath56 and normal vector @xmath42 to the boundary as well as the unit vector @xmath57 .
since the boundary values @xmath49 of the freeform surface only determine the tangential deflection of the rays hitting the boundary , the projection of the law of refraction ( [ eq:4 ] ) on the @xmath50 plane can be used for the calculation of the boundary values.,width=207 ]
we could solve equation ( [ eq:12 ] ) by standard computational fluid dynamic approaches , like finite volume methods , which are appropriate for the numerical treatment of linear advection equations .
based on the nature of equation ( [ eq:16 ] ) , a different approach is proposed in the following . + considering ( [ eq:16 ] )
, we recognize that the boundary values are calculated by the velocity field itself .
this is in contrast to the usual fluid dynamical framework and allows us to separate @xmath39 into an arbitrary number of subareas @xmath58 for each of which we can calculate the boundary values by ( [ eq:16 ] ) .
therefore , the freeform surface can be calculated on each subarea @xmath58 and the solution on @xmath59 by their unification .
this implies that the freeform surface can be constructed by an integration of the omt map along arbitrary paths on @xmath59 , which characterizes the integrability of ray mappings @xcite . thus according to ( [ eq:12 ] ) and ( [ eq:16 ] )
the omt map is approximately integrable as long as ( [ eq:11 ] ) holds .
+ hence , the most convenient way to get the solution of ( [ eq:12 ] ) seems to be a line - by - line integration of ( [ eq:16 ] ) , which along the @xmath60- and @xmath61-direction is equivalent to equation ( [ eq:9 ] ) in a far field approximation .
one possible way of integrating ( [ eq:16 ] ) is shown as an example in fig .
[ fig : integration ] .
thus , only the integration constant of one integral has to be fixed from which the others follow automatically . +
the proposed approach has the useful feature that we do not need to parameterize the boundary @xmath62 , which allows the calculation of freeform surfaces with complex boundary shapes .
the efficiency of the line - by - line integration approach is shown in the next chapter for two challenging design examples . ) by the simple integration of ( [ eq:16 ] ) along straight lines .
first the value @xmath63 is fixed and used for the integration along the green line .
the values of @xmath17 on the green line serve as starting values for the line - by - line integration along the red lines in the orthogonal direction
. then the blue lines are integrated by using the values of @xmath17 on the last red line.,width=207 ]
to show the efficiency of the algorithm , we want to apply it to two design examples . in the first one , we will calculate a freeform lens that maps a collimated beam of uniform intensity on the logo of the institute of applied physics ( iap ) in jena with a resolution of 500 x 500 pixels ( see fig .
[ fig : iap ] ) .
it shows strong intensity gradients between the letters and the background . to omit a division by zero within the implemented omt algorithm @xcite we have to use a background intensity @xmath64 for the input and output intensities . for an appropriate speed of convergence the background intensity
is set to @xmath65 per cent of the maximum intensity .
the second example , which shows smooth intensity variations and a lot of details , is the well - known picture of lena with a resolution of 500 x 500 pixels ( see fig .
[ fig : iap ] ) .
+ since the freeforms were calculated by integrating equation ( [ eq:16 ] ) , the specific characteristics of both pictures do not have any influence on the speed of the lens construction step explained in section [ sec : numerical algorithm ] , but they increase the mapping calculation time . +
the pictures both have a square format .
hence , it is convenient to integrate first along the upper side of the square region and then line - by - line in the orthogonal direction with the starting values given by the first integration .
therefore we have to solve 501 integrals , which took in both cases less than one second in matlab on an intel core i3 at 2.4ghz with 16 gb ram .
this time has to be added up with the mapping calculation time , which strongly depends on the implemented method and the specific features of the picture .
+ the integration constant on the upper left side of the integration area was chosen to be @xmath66 .
values of @xmath67 and @xmath68 were used for the refractive indices , and the source - target distance was chosen to be @xmath69 .
the input and output beam as well as the freeform lens had side length of one .
since every spatial value is normalized the results are scalable .
+ at this point we want to note that according to ( [ eq:16 ] ) the validity of the approximation ( [ eq:11 ] ) can simply be checked by scaling the numerical results with @xmath70 , which of course has to be done for each intensity and configuration individualy . for our examples
the quality of the illumination patterns produced by the raytracing did not change significantly even for distances between the lens and the target plane smaller than the side length of the lens .
+ in both cases the calculated freeform lens was imported as a grid sag surface into zemax to verify our results by a raytracing simulation .
the imported lens data are interpolated by zemax automatically .
the results can be seen in fig .
[ fig : iap ] . +
we presented an efficient numerical method for the single freeform surface design for the shaping of collimated beams .
it is based on the derivation of the condition ( [ eq:10 ] ) for integrable mappings by combining the law of reflaction / refraction and the well - known integrability condition in a suitable way .
we showed that the condition can be fulfilled in the small - angle approximation ( [ eq:11 ] ) by using a ray mapping calculated by optimal mass transport .
equation ( [ eq:11 ] ) therefore represents a quantitative estimate for the applicability of the omt map .
this serves as a theoretical basis for the decoupling of the design process into two separate steps : the calculation of the omt mapping and the construction of the freeform surface with a linear advection equation . on the basis of the finding of appropriate boundary conditions for the advection equation
, we presented a simple numerical algorithm for the surface construction by solving the standard integrals ( [ eq:16 ] ) , which differs from the previously published omt freeform design methods .
+ besides its simplicity , accuracy and quickness , a useful feature of the construction technique is the independence of the freeform boundary shape ( see fig .
[ fig : integration ] ) . by using a proper method for the calculation of the omt mapping ,
this allows for example the calculation of a freeform for disconnected intensities or source and target intensties with two concave boundaries , which is in general a nontrivial problem , but important for applications .
+ the results imply the approximate integrability of the omt map over a wide range of freeform - target plane distances and gives the omt freeform design methods a theoretical basis for collimated input beams .
especially the condition ( [ eq:10 ] ) is interesting , also for other ray mapping techniques than omt methods , since it has to hold for every integrable mapping .
hence , it opens up new possibilities for nearfield calculations as well as generalizations to it for point sources and double freeforms , which are currently under work .
we want to derive equation ( [ eq:8 ] ) from the law of refraction ( [ eq:4 ] ) and the integrability condition ( [ eq:5 ] ) .
since the curl of the incident field @xmath71 vanishs , plugging ( [ eq:4 ] ) into ( [ eq:5 ] ) gives @xmath72 with @xmath73\\ \nonumber\ ] ] it follows @xmath74}_{=[\mathbf{s}_2(\nabla \times \mathbf{s}_2)]\mathbf{s}_2-|\mathbf{s}_2 |^2 ( \nabla \times \mathbf{s}_2 ) } + \mathbf{s}_2 \times [ ( \mathbf{s}_2\nabla)\mathbf{s}_2 ] \}\ ] ] and we can write ( [ eq : a.1 ] ) as @xmath75\mathbf{s}_2 + \mathbf{s}_2 \times [ ( \mathbf{s}_2\nabla)\mathbf{s}_2 ] \right\ } = 0.\ ] ] inserting ( [ eq:4 ] ) and using @xmath76 and @xmath77 leads to @xmath78 + n_1 \{\mathbf{s}_2 \times
[ ( \mathbf{s}_2\nabla)\mathbf{s}_2 ] \}_z = 0\ ] ] and with @xmath79 it follows @xmath80\}_z } { \mathbf{n}\cdot\mathbf{s}_2 } + \mathbf{s}_2(\nabla \times \mathbf{s}_3).\ ] ] using the definition @xmath81 and ( [ eq:6 ] ) the terms in equation ( [ eq : a.5 ] ) can be written as @xmath82 and @xmath83\}_z & = ( \mathbf{s}_2)_x \cdot [ ( \mathbf{s}_2 \cdot \nabla)\mathbf{s}_2]_y - ( \mathbf{s}_2)_y \cdot [ ( \mathbf{s}_2 \cdot \nabla)\mathbf{s}_2]_x \\ & = \begin{pmatrix } -(\mathbf{s}_2)_y \\
( \mathbf{s}_2)_x \end{pmatrix}\cdot \left[(\mathbf{s}_2 \cdot \nabla ) \begin{pmatrix } ( \mathbf{s}_2)_x \\
( \mathbf{s}_2)_y \end{pmatrix } \right ] \\ & = \mathbf{v}\cdot \left[\left(\mathbf{v}^{\perp}\cdot \nabla \right)\mathbf{v}^{\perp}\right ] \end{split}\ ] ] and @xmath84 \\ & = -(z_t - z(x , y))\nabla \mathbf{v}. \end{split}\ ] ]
the authors thank m. esslinger , and m. tessmer for valuable discussions , r. hambach and s. schmidt for valuable discussions and comments on the manuscript , c. liu and d. lokanathan for the help with the zemax implementation and d. musick for the spelling and grammar check .
we also acknowledge the federal ministry of education and research germany for financial support through the project kosimo ( fkz:031pt609x ) .
r. wu , y. zhang , m.m .
sulman , z. zheng , p. bentez , and j.c .
miano , `` initial design with l2 monge - kantorovich theory for the monge ampre equation method in freeform surface illumination design , '' opt . express * 22*(13 ) , 1616116177 ( 2014 ) v.i .
oliker , `` mathematical aspects of design of beam shaping surfaces in geometrical optics , '' in _ trends in nonlinear analysis _ , m. kirkilionis , s. kromker , r. rannacher , and f. tomi , eds .
( springer - verlag , 2003 ) , pp .
193 - 222 .
v. oliker , and b. cherkasskiy , `` controlling light with freeform optics : recent progress in computational methods for optical design of freeform lenses with prescribed irradiance properties , '' proc .
spie * 9191 * , 919105919105 - 7 ( 2014 ) . | the efficient design of continuous freeform surfaces , which transform a given source into an arbitrary target intensity , remains a challenging problem . a popular approach
are ray - mapping methods , where first a ray mapping between the intensities is calculated and in a subsequent step the surface is constructed .
the challenging part hereby is the to find an _ integrable _ mapping ensuring a _ continuous _ surface .
based on the law of reflection / refraction and the well - known integrability condition , we derive a general condition for the surface and ray mapping for a collimated input beam . it is shown that in a small - angle approximation a proper mapping can be calculated via optimal mass transport .
we show that the surface can be constructed by solving a linear advection equation with appropriate boundary conditions .
the results imply that the optimal mass transport mapping is approximately integrable over a wide range of distances between the freeform and the target plane and offer an efficient way to construct the surface by solving standard integrals .
the efficiency is demonstrated by applying it to two challenging design examples .
pacs numbers : : ` 42.15.-i , 42.15.eq ` . |
high energy ion accelerators , transport systems and storage rings @xcite have a wide range of applications ranging from basic research in high energy and nuclear physics , to applications such as heavy ion fusion , spallation neutron sources , and nuclear waste transmutation
. of particular importance at the high beam currents and charge densities of interest for heavy ion fusion are the effects of the intense self - fields produced by the beam space charge and current on determining detailed equilibrium , stability , and transport properties . in general
, a complete description of collective processes in intense charged particle beams is provided by the nonlinear vlasov - maxwell equations @xcite for the self - consistent evolution of the beam distribution function , @xmath0 , and the self - generated electric and magnetic fields , @xmath1 and @xmath2 .
while considerable progress has been made in analytical and numerical simulation studies of intense beam propagation @xcite , the effects of finite geometry and intense self - fields often make it difficult to obtain detailed predictions of beam equilibrium , stability , and transport properties based on the vlasov - maxwell equations . to overcome this complexity ,
considerable theoretical progress has also been made in the development and application of one - dimensional vlasov - maxwell models @xcite to describe the longitudinal beam dynamics for a long coasting beam , with applications ranging from plasma echo excitations , to the investigation of coherent soliton structures , both compressional and rarefactive ( hole - like ) .
such one - dimensional vlasov descriptions rely on using a geometric - factor ( @xmath3-factor ) model @xcite to incorporate the average effects of the transverse beam geometry and the surrounding wall structure . despite the many successful applications of such one - dimensional vlasov models to describe the longitudinal dynamics of long costing beams , even a one - dimensional vlasov description is often too complicated to be analyzed in detail , except in some simple limiting cases .
the next level of description of the nonlinear beam dynamics is provided by the macroscopic fluid equations , which correspond to the first three momentum moments of the vlasov equation , with some particular closure scheme which relates the higher moments to the first three @xcite .
the usual closure assumption for collisionless plasma is provided by the adiabatic equation of state ( @xmath4 , where @xmath5 is the entropy per the unit volume ) , which expresses the thermal pressure as a function of the density . in one dimension , this is given by @xmath6 @xcite , where @xmath7 is the line density of beam particles and @xmath8 is the line pressure .
such a one - dimensional fluid model , combined with the adiabatic equation of state and a g - factor model for the average electric field , fits in the class of one - dimensional fluid problems which can be solved exactly @xcite using the formalism described in sec .
ii . in presently envisioned configurations for heavy ion fusion , multiple , high - current ,
heavy ion beams are focused to a small spot size onto the target capsule . to achieve a high - intensity beam focused onto the target ,
the beams are first accelerated and then compressed longitudinally .
one of the possible ways to compress the beam longitudinally is to use a the drift compression scheme illustrated in fig .
[ fig1 ] @xcite .
the scheme consists of two parts . in the first stage ,
an initial tilt in the longitudinal velocity profile @xmath9 is imposed on the long charge bunch with some particular line density profile @xmath10 .
after time @xmath11 , the beam line density evolves to a profile @xmath12 , with velocity profile @xmath13 . at this time , an additional velocity tilt @xmath14 is imposed on the beam , and the beam is left with line density @xmath12 and velocity tilt @xmath15 .
we refer to this stage as the _ beam shaping _ stage .
this stage requires beam manipulation ( imposing the velocity tilt ) and is done when the charge bunch is very long . at this stage ,
the longitudinal pressure and electric field are negligible , and the beam dynamics is governed by free convection . during this stage the beam may or may not be compressed .
the purpose of this stage is to shape the beam line density profile into a certain intermediate line density profile @xmath12 with velocity profile @xmath15 , such that during the next stage , which we refer to as the _ drift compression _ stage , after further axial drift during the time interval @xmath16 ( see fig . [ fig1 ] ) , the beam is compressed longitudinally until the space - charge force or the internal thermal pressure stops the longitudinal compression of the charge bunch @xcite . at the point of maximum compression , the velocity tilt profile is completely removed and the beam is left with desired line density @xmath17 .
the final focus magnets then focus the beam onto the target , and the beam heats and compresses the target fuel .
stated this way , the longitudinal drift compression problem in the beam frame is equivalent to the time - reversed problem of the beam expanding into vacuum with zero initial velocity profile , and the specified initial line density profile @xmath17 which is desired at the time of maximum compression before final focusing . in this paper we employ a one - dimensional warm - fluid model with adiabatic equation of state to study this problem analytically for arbitrary final ( maximally compressed ) line density profiles @xmath17 .
we consider here the two separate cases corresponding to a cold beam , and a pressure - dominated beam . in the case of a cold beam ,
the internal thermal pressure is negligible , and the dynamics of the drift compression is governed by the self - generated electric field . in the case of a pressure - dominated beam ,
the self - generated electric field is negligible , and the beam compresses under the influence of the thermal pressure and the initial velocity tilt @xmath15 .
one of the compression scenarios considered for heavy ion fusion is _ neutralized _ drift compression , where the beam propagates through a charge - neutralizing background plasma as it compresses longitudinally .
in such a scenario , the beam is compressed only against the internal pressure . for simplicity ,
the present analysis is carried out in the beam frame where the particle motions are nonrelativistic .
the final results can be then lorentz transformed back to the laboratory frame , moving with axial velocity @xmath18 relative to the average motion of the particles in the beam frame .
this paper is organized as follows . in sec .
ii , we briefly describe the one - dimensional warm - fluid model equations and the formalism used for solving them analytically . in secs .
iii and iv , the general solution for the expansion problem ( the inverse to the _ drift compression stage _
problem ) is obtained analytically for the two cases corresponding to a cold beam , and a pressure - dominated beam . in sec .
v the general solution to drift compression stage obtained in secs .
iii and iv is illustrated by several examples . in sec .
vi we briefly discuss the _ beam shaping _ stage .
in the present analysis , we employ a one - dimensional warm - fluid model @xcite to describe the longitudinal nonlinear beam dynamics with average electric field given by the g - factor model with @xmath19 @xcite .
for example , for a space - charge - dominated beam with flat - top density profile in the transverse plane , @xmath20 @xcite . here , @xmath21 is the line density , @xmath22 is the charge of a beam particle , @xmath23 is the conducting wall radius , and @xmath24 is the beam radius .
generally , the beam radius , and therefore the @xmath3-factor , are functions of the line density and the external transverse focusing , and can change during the beam compression . in most of the drift compression scenarios it is preferable to maintain the beam radius ( and therefore the g - factor ) constant during the beam compression by adjusting the external transverse focusing @xcite .
therefore , in the present analysis , we assume that the g - factor is a constant .
the macroscopic fluid equations for the line density @xmath21 , the average longitudinal beam velocity @xmath25 , and the longitudinal line pressure @xmath26 are given by @xcite @xmath27 where @xmath28 for a triple - adiabatic equation of state .
here we have introduced the effective potential @xmath29 defined by @xmath30 where @xmath31 and @xmath32 are constants with dimensions of @xmath33 , @xmath34 is the mass of a beam particle , and @xmath35 and @xmath36 are constants with the dimensions of line density and line pressure , respectively . for future application in secs .
iii - vi , in the remainder of this section we summarize well - established theoretical technique developed in fluid mechanics @xcite that can be used to solve the nonlinear fluid equations ( [ eq1 ] ) and ( [ eq2 ] ) . by introducing the velocity potential @xmath37 , where @xmath38
, we can rewrite eq .
( [ eq2 ] ) as @xmath39 the full differential of @xmath37 then becomes @xmath40[eq4]next , following landau and lifshitz @xcite , we introduce the legendre transform @xmath41+td\left({v^2\over 2 } + w \right).\label{eq5}\ ] ] introducing @xmath42 , eq .
( [ eq5 ] ) can be expressed as @xmath43 it follows from eq .
( [ eq6 ] ) that @xmath44 can be considered as a function of the new independent variables @xmath45 , and that @xmath46 therefore , if the function @xmath44 is known as a function of its arguments @xmath45 , then eq . ( [ eq7 ] ) gives @xmath45 as implicit functions of @xmath47 . to obtain the equation for @xmath44 we rewrite eq .
( [ eq1 ] ) as @xmath48 where @xmath49 assuming that @xmath50 is not a definite function of @xmath29 [ @xmath51 , we multiply eq .
( [ eq8 ] ) by @xmath52 and use the multiplication property for determinants .
this gives @xmath53 the case when @xmath50 is a definite function of @xmath29 in some region of the @xmath47 plane corresponds to _ a simple wave _ and will be considered later . because @xmath54 , eq .
( [ eq10 ] ) reduces to @xmath55 substituting eq .
( [ eq7 ] ) into eq .
( [ eq11 ] ) , we obtain the equation for @xmath44 @xcite @xmath56 note that eq .
( [ eq12 ] ) is a linear partial differential equation for the function @xmath57 . by introducing the effective sound speed defined by @xmath58 ,
we can rewrite eq .
( [ eq12 ] ) as @xmath59 where @xmath60 is to be regarded as a function of @xmath29 .
equation([eq13 ] ) together with eq .
( [ eq7 ] ) can be used to obtain the solution to the system of equations ( [ eq1 ] ) and ( [ eq2 ] ) everywhere in the @xmath47 plane except in the regions corresponding to _ simple wave _ solutions where @xmath61 @xcite . in this case
, the jacobian @xmath62 vanishes identically . in deriving eq .
( [ eq10 ] ) , we divided eq .
( [ eq8 ] ) by this jacobian , and the solution for which @xmath63 is not recovered .
thus , a simple wave solution can not be recovered from the general equation ( [ eq12 ] ) . if @xmath50 is a function of @xmath7 only , as in a simple wave , we can rewrite eqs.([eq1 ] ) and ( [ eq2 ] ) as @xcite @xmath64 since @xmath65 we obtain from eqs .
( [ eq21 ] ) and ( [ eq22 ] ) @xmath66 however , because @xmath67 , it follows that @xmath68 , so that @xmath69 , and therefore @xmath70 next , we combine eqs .
( [ eq23 ] ) , ( [ eq24 ] ) and ( [ eq25 ] ) to give @xmath71 .
integrating with respect to @xmath72 then gives @xcite @xmath73+f(v),\label{eq28}\ ] ] where @xmath74 is an arbitrary function of the velocity @xmath50 determined from the initial conditions , and @xmath75 is given by eq .
( [ eq25 ] ) .
equations ( [ eq25 ] ) and ( [ eq28 ] ) give the general solution for the _ simple wave_. the two signs in eqs .
( [ eq25 ] ) and ( [ eq28 ] ) correspond to the direction of wave propagation : @xmath76 is for a wave propagating in the positive @xmath77 direction , and @xmath78 is for a wave propagating in the negative @xmath77 direction .
it is also convenient to solve eqs .
( [ eq1 ] ) and ( [ eq2 ] ) using the method of _ characteristics _ @xcite .
multiplying eq .
( [ eq1 ] ) by @xmath79 , and then adding and subtracting from eq .
( [ eq2 ] ) , and making use of the relation @xmath80 , we obtain @xmath81 we now introduce the new unknown functions @xmath82 which are called _
riemann invariants_. in terms of @xmath83 and @xmath84 , the equations of motion take the simple form @xcite @xmath85j_+=0,\ \ \ \ \ \left[{\partial\over \partial t}+(v - c){\partial\over \partial x}\right]j_-=0.\label{eq33}\ ] ] the differential operators acting on @xmath83 and @xmath84 are the operators for differentiation along the curves @xmath86 and @xmath87 ( called _ characteristics _ ) in the @xmath47 plane given by the equations @xmath88 the values of @xmath50 and @xmath89 at every point of the @xmath47 plane are given by the values of the riemann invariants @xmath83 and @xmath84 which are transported to this point along the @xmath86 and @xmath87 characteristics from the region where the values of @xmath83 and @xmath84 ( and therefore @xmath50 and @xmath89 ) are known .
equations([eq30 ] ) and ( [ eq34 ] ) are very convenient for numerical solution of the equations of motion and also for general analysis of the flow .
the totality of available space - time generally consists of the regions where either a simple wave solution [ eq .
( [ eq28 ] ) ] or a general solution [ solution to eq .
( [ eq13 ] ) together with eq .
( [ eq7 ] ) ] is applicable .
the boundary between the simple wave solution and the general solution , like any boundary between two analytically different solutions , is a characteristic @xcite . in solving particular problems ( see sec .
v ) , the value of the function @xmath57 on this boundary characteristic must be determined .
the matching condition at the boundary between the simple wave solution and the general solution is obtained by substituting eq .
( [ eq7 ] ) for @xmath77 and @xmath72 into the equation for the simple wave [ eq . ( [ eq28 ] ) ] .
this gives @xmath90 moreover , in the simple wave solution ( and therefore on the boundary characteristic ) , we obtain @xmath91 . substituting into eq .
( [ eq35 ] ) then gives @xmath92 equation ( [ eq36 ] ) can be integrated to give @xcite @xmath93 which determines the required boundary value of @xmath44 .
here we consider two separate cases . for the case of a cold beam ( @xmath94 and @xmath95 ) it follows that @xmath96 [ eq . ( [ eq2a ] ) ] , and @xmath97 . in the opposite limit where we can neglect the electric field compared to the thermal pressure
, it follows that @xmath98 , and @xmath99 . introducing in place of @xmath29 the variable @xmath100 , where @xmath101 , we can rewrite eq.([eq13 ] ) as @xmath102 equation ( [ eq15 ] ) is an ordinary wave equation whose general solution is @xmath103 where @xmath104 and @xmath105 are arbitrary functions . to find the general solution to eq .
( [ eq14 ] ) we first fourier transform with respect to the @xmath50 dependence .
this gives @xmath106 equation ( [ eq17 ] ) is bessel s equation of order zero which has two hankel functions , @xmath107 and @xmath108^{*}$ ] , as independent solutions . here
@xmath109 represents complex conjugate . using the integral representation of the hankel function @xcite , @xmath110 the general solution to eq .
( [ eq14 ] ) can be expressed as @xmath111 where @xmath112 and @xmath113 are arbitrary functions .
finally , we can rewrite eq .
( [ eq19 ] ) as @xmath114,\label{eq20}\end{aligned}\ ] ] where @xmath104 and @xmath105 are arbitrary functions such that the integrals in eq .
( [ eq20 ] ) converge .
equation ( [ eq20 ] ) provides the general solution to eq .
( [ eq14 ] ) . in the regions pertaining to the simple wave solution , eq.([eq25 ] ) gives the general relation between velocity and the density or sound speed in the wave .
in the two cases considered here , @xmath115 for a cold beam @xmath116 , and @xmath117 for a pressure - dominated beam @xmath118 .
for these two cases @xmath119 where @xmath120 is a constant . equations ( [ eq26 ] ) and ( [ eq27 ] ) together with eq .
( [ eq28 ] ) give a simple wave solution for the two cases considered in this section . from eq .
( [ eq30 ] ) , the corresponding riemann invariants can be expressed as @xmath121
in this section we make use of eqs .
( [ eq16 ] ) and ( [ eq20 ] ) to solve the initial value problem for the case of beam expansion into vacuum .
the initial conditions for this problem are zero flow velocity at every point , @xmath122 , and prescribed density profile @xmath123 , which expresses the initial line density as a function of @xmath77 . at some later time
@xmath124 , the density and velocity profiles will be given by the functions @xmath125 and @xmath126 which are the solutions to eqs.([eq1 ] ) and ( [ eq2 ] ) .
since the equations of motion [ eqs.([eq1 ] ) and ( [ eq2 ] ) ] are time - reversible , the flow described by @xmath127 and @xmath128 are also solutions to these equations with initial conditions @xmath129 and @xmath130 . at time @xmath124
this flow has zero velocity profile ( @xmath131 ) and the density profile is given by the initial profile for the expansion problem , i.e. , @xmath132 . to solve the initial value problem
we assume that the density profile @xmath133 , or equivalently the sound velocity profile @xmath134 , decreases monotonically to zero at the beam boundary @xmath135 , is an even function of @xmath77 , and is an invertable function for @xmath136 everywhere where the density is non - zero .
therefore , we assume that at @xmath137 the inverted profile @xmath138 is known .
the condition that @xmath134 decreases monotonically to zero at the beam boundary means that no rarefaction wave is launched from the boundary into the beam as it expands .
we will treat the case with discontinues in @xmath134 at the beam boundary in one of the examples in sec .
v. since we are interested in the time - reversed problem of beam compression , we assume that multi - valued flow does not form as the beam expands .
this is equivalent to considering only initial density profiles with first derivative decreasing continuously from the beam center to the beam edge .
this guarantees that the portions of the beam with smaller density accelerate faster than the portions with lower density , and as a result , the flow is never multi - valued .
we will treat the case with multi - valued flow in one of the examples in sec . v. the region of flow in the @xmath47 plane and its boundaries
are illustrated in fig .
it is obvious that the flow is symmetric under reflection , @xmath139 , and therefore we need only to solve the equations in the region @xmath136
. in general , there are four regions of flow .
( see figs .
[ fig2 ] and [ fig3 ] . ) each is separated from the others by two characteristics , the @xmath87 characteristic @xmath140 on which @xmath141 , and the @xmath86 characteristic @xmath142 on which @xmath143 .
the boundary conditions are given by @xmath144 is the initial beam half - width , and @xmath145 is the coordinate of the beam edge . as evident from fig .
[ fig2 ] , the flow at every point in region i is brought to this point by the characteristics originating from the x - axis at @xmath137 . hence , the flow in this region is fully determined by the boundary conditions @xmath146 and @xmath147 .
the flow at every point in the region ii , which is separated from region i by the @xmath86 characteristic q originating from the origin in the @xmath47 plane , is brought to this point by the characteristics originating from the @xmath148 line where @xmath149 , and from region i. therefore , the boundary condition for region ii is given by @xmath149 and by the flow on the separating characteristic q. the flow at every point in region iii , which is adjacent to the beam edge and separated from region i by the @xmath87 characteristic p originating from the beam edge at @xmath137 in the @xmath47 plane , is brought to this point by the characteristics originating from the beam edge where @xmath150=0 $ ] , and from region i. therefore , the boundary condition for region iii is given by @xmath150=0 $ ] and by the flow on the separating characteristic p. the flow at every point in region iv , which is separated from region ii by the p characteristic and from region iii by the q characteristic , is brought to this point by the characteristics originating from region ii and region iii .
therefore the boundary condition for region iv is given by the flow on the separating characteristics p and q. the function @xmath151 and eq .
( [ eq7 ] ) provide the map of the flow region in the @xmath47 plane illustrated in fig.[fig2 ] , to the @xmath152 plane ( fig .
[ fig3 ] ) .
the region is a triangle @xmath153 limited from above by the @xmath86 characteristic q which is a straight line in the @xmath152 plane since on this characteristic @xmath154 ( @xmath101 ) .
this mapping is not one - to - one .
in fact regions i and ii and regions iii and iv in the @xmath47 plane map into the same regions in the @xmath152 plane , which means that in the @xmath152 plane there will be four functions , @xmath155 , @xmath156 , @xmath157 and @xmath158 , defined inside the area depicted in fig .
[ fig3 ] , which map the depicted @xmath152 region back into regions i , ii , iii and iv in the @xmath47 plane , respectively , by means of eq.([eq7 ] ) . since at @xmath137 , @xmath147 and @xmath159 , by making use of eq .
( [ eq7 ] ) we obtain the boundary conditions for @xmath160 in the @xmath152 plane in region i , which can be expressed as @xmath161 since @xmath149 at @xmath148 , the boundary condition for @xmath162 is @xmath163 the second boundary condition for @xmath162 reflects continuity of the mapping in eq .
( [ eq7 ] ) , @xmath164 by making use use of @xmath150=0 $ ] , the definition @xmath165 and eq .
( [ eq7 ] ) , we obtain the boundary condition for @xmath166 , @xmath167 the second boundary condition for @xmath166 reflects continuity of the mapping in eq .
( [ eq7 ] ) , @xmath168 finally , the boundary conditions for @xmath169 reflects continuity of the mapping in eq .
( [ eq7 ] ) , @xmath170 and @xmath171 next , we consider separately the two cases corresponding to a cold beam , and a pressure - dominated beam .
the general solution for the case of a pressure - dominated beam is given by eq .
( [ eq16 ] ) . to satisfy the boundary condition in eq .
( [ eq39 ] ) , we are required to choose @xmath172 substituting eq .
( [ eq44 ] ) into eq .
( [ eq40 ] ) , we obtain @xmath173 , and therefore @xmath174 here , we have chosen the integration constant so that @xmath175 . substituting eq .
( [ eq45 ] ) into eq .
( [ eq44 ] ) then gives @xmath176 in region ii , to satisfy the boundary conditions in eq.([eq41 ] ) , we are required to choose @xmath177 to satisfy the boundary conditions in eqs .
( [ eq42 ] ) and ( [ eq43 ] ) we choose @xmath178 .
hence , the solution in region ii is given by @xmath179 it is readily shown that the solution in region iii which satisfies all of the boundary conditions in eqs .
( [ eq43a ] ) , ( [ eq43b ] ) and ( [ eq43c ] ) is given by @xmath180 and the solution in region iv which satisfies all of the boundary conditions in eqs .
( [ eq43d ] ) , ( [ eq43e ] ) , ( [ eq43f ] ) and ( [ eq43 g ] ) is given by @xmath181 finally , using eq .
( [ eq7 ] ) and the definition @xmath182 , we obtain the solutions in region i , @xmath183 , \nonumber \\ & & t={1\over 2c}[x_0(c - v)-x_0(c+v)],\label{eq49}\end{aligned}\ ] ] in region ii , @xmath184 , \nonumber \\ & & t={1\over 2c}[x_0(c - v)+x_0(c+v)].\label{eq50}\end{aligned}\ ] ] in region iii , @xmath185 , \nonumber \\ & & t={1\over 2c}[x_0(v - c)-x_0(c+v)],\label{eq51}\end{aligned}\ ] ] and in region iv , @xmath186 , \nonumber \\ & & t={1\over 2c}[x_0(v - c)+x_0(c+v)].\label{eq52}\end{aligned}\ ] ] equations ( [ eq49 ] ) and ( [ eq50 ] ) give the implicit solution describing the expansion of a pressure - dominated beam .
we can also obtain the formulas for the asymptotic solution as @xmath187 or @xmath188 . indeed , for @xmath187 or @xmath188 the flow is almost entirely in region iv .
using eq .
( [ eq52 ] ) , we obtain @xmath189 finally , in the leading approximation , we can rewrite eq.([eq52a ] ) as @xmath190 evidently , the density profile given by eq .
( [ eq52c ] ) is correctly normalized .
the same solution can be also obtained from a kinetic description .
it s been shown in ref .
@xcite that eqs .
( [ eq1 ] ) and ( [ eq2 ] ) [ together with the adiabatic pressure relation @xmath191 are the two key moments of the kinetic vlasov equation for a waterbag distribution function ( @xmath192 in an enclosed area of phase space ) . indeed
if we denote the upper curve in fig .
[ fig4 ] as @xmath193 and the lower curve as @xmath194 , than by multiplying the vlasov equation for @xmath195 @xmath196 by @xmath197 and by @xmath198 , integrating over @xmath198 , and keeping in mind that @xmath199 inside the region limited from above by @xmath193 and from below by @xmath194 , we obtain @xmath200+{1\over 2}{\partial \over \partial x}[v_+(x , t)^2-v_-(x , t)^2]=0,\label{eq54}\\ & & { 1\over 2}{\partial \over
\partial t}[v_+(x , t)^2-v_-(x , t)^2]+{1\over 3}{\partial \over \partial x}[v_+(x , t)^3-v_-(x , t)^3]=0.\label{eq55}\end{aligned}\ ] ] introducing the line density and flow velocity defined by @xmath201}[v_+(x , t)-v_-(x , t)],\label{eq56}\\ & & v(x , t)={1\over 2}[v_+(x , t)-v_-(x , t)],\label{eq57}\end{aligned}\ ] ] where @xmath35 is the density at @xmath148 at @xmath137 , we can rewrite the eqs . ( [ eq54 ] ) and ( [ eq55 ] ) in familiar form @xmath202=0,\label{eq58}\\ & & { \partial\over \partial t}[\lambda(x , t)v(x , t)]+{\partial \over \partial x}[\lambda(x , t)v(x , t)^2]+{1\over 12 } { [ v_+(0,0)-v_-(0,0)]^2\over \lambda_0 ^ 2}{\partial \over \partial x}\lambda(x , t)^3=0.\label{eq59}\end{aligned}\ ] ] comparing with eq .
( [ eq2 ] ) , we obtain @xmath203 ^ 2/ 4\lambda_0 ^ 2)\lambda^2 $ ] , or @xmath204/ 2\lambda_0)\lambda(x , t)=(1/2)[v_+(x , t)-v_-(x , t)]$ ] . therefore , @xmath205 and @xmath206 . if the initial profiles are given by @xmath147 and @xmath146 , then @xmath207 and @xmath208 .
since eq .
( [ eq53 ] ) represents the free - streaming motion of the particles in phase space along straight - line trajectories , we readily obtain the expressions for @xmath194 and @xmath193 , @xmath209,\label{eq60}\\ & & v_-(x , t)=\pm c_0[x - v_-(x , t)t].\label{eq61}\end{aligned}\ ] ] here , the @xmath210 sign in eq .
( [ eq61 ] ) holds for @xmath211 ( @xmath212 is the coordinate of the beam edge at @xmath137 ) and corresponds to regions i and ii in fig .
[ fig2 ] , and the @xmath213 sign holds for @xmath214 and corresponds to regions iii and iv in fig.[fig2 ] ( see fig .
[ fig4 ] ) . equations ( [ eq60 ] ) and ( [ eq61 ] ) can be rewritten as equations for @xmath215 and @xmath25 , @xmath216\pm c_0[x-(v - c)t]\right\},\label{eq62}\\ & & v(x , t)={1\over 2}\left\{c_0[x-(v+c)t]\mp c_0[x-(v - c)t]\right\}.\label{eq63}\end{aligned}\ ] ] by adding and subtracting eqs .
( [ eq62 ] ) and ( [ eq63 ] ) , and inverting the resulting equations , we obtain @xmath217 for regions i and ii , and @xmath218 for regions iii and iv . the @xmath213 sign in eqs.([eq64])-([eq67 ] ) corresponds to region i [ eqs.([eq64 ] ) and ( [ eq65 ] ) ] and region iii [ eqs . ( [ eq66 ] ) and ( [ eq67 ] ) ] , and the @xmath210 sign corresponds to region ii [ eqs .
( [ eq64 ] ) and ( [ eq65 ] ) ] and region iv [ eqs.([eq66 ] ) and ( [ eq67 ] ) ] ( see figs . [ fig2 ] and [ fig4 ] ) .
the @xmath219 signs appear here because we have assumed an even initial profile @xmath220 .
finally , by adding and subtracting , eqs.([eq64])-([eq67 ] ) take the form shown in eqs.([eq49])-([eq52 ] ) .
here we use the general solution in eq .
( [ eq20 ] ) to solve the same initial value problem as discussed in the previous section , applied now to the case of a cold beam . to satisfy the boundary condition in eq .
( [ eq39 ] ) we are required to choose @xmath221.\label{eq68}\ ] ] substituting eq .
( [ eq68 ] ) into eq .
( [ eq40 ] ) , we obtain @xmath222 equation ( [ eq69 ] ) can be inverted by using the integral abel transform in appendix a. this gives @xmath223 where @xmath224 , @xmath225 is the heaviside step - function , and @xmath226 .
note from eq .
( [ eq68 ] ) that in regions i and ii , where @xmath227 , the argument of @xmath228 in eq.([eq68 ] ) is positive , and we can use the form of @xmath228 defined in eq .
( [ eq70 ] ) . for @xmath229 ( regions iii and iv ) , the argument of the function under the integral in the first term in general solution in eq .
( [ eq20 ] ) can become negative .
next , we show that the solution of the form in eq .
( [ eq68 ] ) with @xmath230 continued to the regions where @xmath231 as @xmath232 , or @xmath233 , will satisfy the boundary conditions in eqs.([eq43a])-([eq43c ] ) . indeed , by expanding @xmath228 in a taylor series in eq .
( [ eq68 ] ) for @xmath188 , we obtain @xmath234\ln\left({c_0+v/2\over c}\right)+o(c^2),\label{eq70a}\ ] ] and therefore @xmath235.\label{eq70b}\ ] ] differentiating with respect to @xmath89 , and taking the limit @xmath188 in eq .
( [ eq70b ] ) , we obtain @xmath236\over c}+o\left[c\ln\left({c_0\over c } \right)\right]=0,\label{eq70bb}\end{aligned}\ ] ] provided @xmath237 .
it readily follows that the continuity conditions in eqs .
( [ eq43b])-([eq43c ] ) are also satisfied .
therefore , the solutions in regions i and iii are given by @xmath238.\label{eq70c}\ ] ] where @xmath239 to obtain the solutions in regions ii and iv , we note that the function @xmath240.\label{eq70e}\ ] ] satisfies the condition in eq .
( [ eq41 ] ) .
also , if we choose @xmath228 as in eq .
( [ eq70d ] ) , the second term ( and its first derivatives ) in both equations ( [ eq68 ] ) and ( [ eq70e ] ) is zero on the dividing characteristic @xmath241 , and therefore all of the continuity conditions in eqs.([eq43d])-([eq43 g ] ) are also satisfied .
equations ( [ eq70c])-([eq70e ] ) together with eq .
( [ eq7 ] ) give the formal solution of the expansion problem for the case of a cold beam . finally , substituting eq .
( [ eq70d ] ) into eqs.([eq70c ] ) and ( [ eq70e ] ) , changing the order of integration , and performing the integrations , we obtain @xmath242,\label{eq71 } \\
\chi^{ii}(v , c)=&-&{2c\over \pi}\int_{1-v/2c}^{1+v/2c}dq\sqrt{q}x_0(c q)k\left[{(v/2c)^2-(q-1)^2\over 4q}\right]\nonumber\\ & -&{8c\over \pi}\int_{1+v/2c}^{c_0/c}{dqqx_0(c q)\over \sqrt{(q+1)^2-(v/2c)^2}}k\left[{(q-1)^2-(v/2c)^2\over ( q+1)^2-(v/2c)^2 } \right],\label{eq72 } \\
\chi^{iii}(v , c)=&-&{2c\over \pi}\int_{v/2c-1}^{1+v/2c}dq\sqrt{q}x_0(c q)k\left[{(v/2c)^2-(q-1)^2\over 4q}\right]\nonumber\\ & -&{4c\over \pi}\int_{0}^{v/2c-1}{dqqx_0(c q)\over \sqrt{(v/2c)^2-(q-1)^2}}k\left[{4q\over ( v/2c)^2-(q-1)^2}\right],\label{eq73 } \\
\chi^{iv}(v , c)=&-&{2c\over \pi}\int_{v/2c-1}^{1+v/2c}dq\sqrt{q}x_0(c q)k\left[{(v/2c)^2-(q-1)^2\over 4q}\right]\nonumber\\ & -&{4c\over \pi}\int_{0}^{v/2c-1}{dqqx_0(c q)\over \sqrt{(v/2c)^2-(q-1)^2}}k\left[{4q\over ( v/2c)^2-(q-1)^2}\right ] \nonumber \\ & -&{8c\over \pi}\int_{1+v/2c}^{c_0/c}{dqqx_0(c q)\over \sqrt{(q+1)^2-(v/2c)^2}}k\left[{(q-1)^2-(v/2c)^2\over ( q+1)^2-(v/2c)^2 } \right ] .
\label{eq74}\end{aligned}\ ] ] here , @xmath243 is the complete elliptic integral of the first kind @xcite . in sec .
v , we illustrate the application of these solutions with several examples .
we can also obtain the formulas for the asymptotic solution as @xmath187 or @xmath188 .
indeed , for @xmath187 or @xmath188 , the flow is almost entirely in regions ii and iv . using eqs .
( [ eq70a ] ) and ( [ eq70e ] ) , we obtain @xmath244\ln\left({c_0 ^ 2-(v/2)^2\over c^2}\right)+o(c^2),\label{eq74a}\ ] ] and therefore @xmath245 finally , in the leading approximation , we can rewrite eq.([eq74c ] ) as @xmath246 where @xmath247 and @xmath248 here @xmath249 is the scaled initial line density profile .
one can readily verify that the density profile given by eq .
( [ eq74d ] ) is correctly normalized , and that @xmath250 .
in this section , we apply the formalism developed in sec .
ii - iv to several examples with different initial density profiles . as a first example
, we consider here the case of an initial parabolic density profile for @xmath123 with @xmath251\theta(|x|<x_0).\label{eq75a}\ ] ] for a cold beam , @xmath252 .
substituting eq.([eq75a ] ) into eq .
( [ eq74e ] ) and integrating , we obtain @xmath253\theta(z < c_0).\label{eq76}\ ] ] next we substitute eq .
( [ eq76 ] ) into the integral @xmath254 , \label{eq77}\end{aligned}\ ] ] and define @xmath255 where we have introduced new variables @xmath256 in term of the new variables , it follows that @xmath257,\label{eq80}\\ & & \chi^{ii,\ iv}(a , b)=-i^+(a , b)-i^-(a , b)=\nonumber \\ & & -{c_0x_0\over ( a+b)^2}\left[(a-2b)\sqrt{a^2 - 1}+(b-2a)\sqrt{y^2 - 1}+(2ab-1)\ln{a+\sqrt{a^2 - 1}\over b-\sqrt{b^2 - 1}}\right],\label{eq81}\end{aligned}\ ] ] by introducing the scaled variables @xmath258 , @xmath259 and @xmath260 , we can rewrite eq .
( [ eq7 ] ) as @xmath261\bar\chi,\label{eq82}\\ & & \bar x=-{(a^2-b^2)\over 4}\left[a{\partial\over\partial a}+b{\partial\over\partial b}\right]\bar\chi-{(a+b)\over 4}\left[{\partial\over\partial b}-{\partial\over\partial a}\right]\bar\chi.\label{eq83}\end{aligned}\ ] ] finally , substituting eqs .
( [ eq80 ] ) and ( [ eq81 ] ) into eqs .
( [ eq82 ] ) and ( [ eq83 ] ) , and using eqs .
( [ eq79 ] ) , we obtain after some lengthy algebra the solution in regions i and iii , @xmath262 and in regions ii and iv , @xmath263 here we have introduced @xmath264 and @xmath265
. equations ( [ eq84])-([eq87 ] ) can be easily inverted .
the result is @xmath266 ^ 2\left({2f\over 1+f^2}\right)^4},\label{eq89}\\ & & \left({v\over 1-f^2 } \right)^2+\left({c\over f}\right)^2=\left({2c_0\over 1+f^2}\right)^2,\label{eq90}\\ & & { \lambda\over\lambda_0}=\left({2f\over 1+f^2}\right)^2\left[1-\left({x\over x_0}\right)^2\left({2f\over 1+f^2}\right)^4\right],\label{eq91}\end{aligned}\ ] ] where @xmath267 is the solution of the transcendental equation @xmath268 the solutions in eqs .
( [ eq88])([eq92 ] ) describe the familiar self - similar solution @xcite for a parabolic density profile and is plotted in fig .
[ fig5 ] . using eqs .
( [ eq74d ] ) and ( [ eq76 ] ) , we obtain the asymptotic solution as @xmath187 @xmath269,\ \ \ \ v(x , t)={x\over t},\ \ \text{for}\ \ t\rightarrow \infty . \label{eq91a}\end{aligned}\ ] ] the exact solution given by eq .
( [ eq91 ] ) and asymptotic solution given by eq .
( [ eq91a ] ) are compared in fig.[fig6 ] ( line b ) for @xmath270 . for a pressure - dominated beam , @xmath271 and threrefore the initial profile for @xmath138
is given by @xmath272 .
using eq .
( [ eq75a ] ) and eqs .
( [ eq62 ] ) and ( [ eq63 ] ) we obtain the implicit solution in regions i and ii , @xmath273 ^ 2,\label{eq142}\\ & & \bar c-\bar v=1-[\bar x-(\bar v-\bar c)\bar t]^2,\label{eq143}\end{aligned}\ ] ] and in regions iii and iv , @xmath273 ^ 2,\label{eq144}\\ & & \bar v-\bar c=1-[\bar x-(\bar v-\bar c)\bar t]^2.\label{eq145}\end{aligned}\ ] ] solving eqs .
( [ eq142])([eq145 ] ) for @xmath274 and @xmath275 , we obtain @xmath276,\label{eq146 } \\ & & \bar v={\bar x \over \bar t}+{1\over 4\bar t^2}[\sqrt{1 + 4\bar t^2 - 4\bar x \bar t}-\sqrt{1 + 4\bar t^2 + 4\bar x \bar t}],\label{eq147}\end{aligned}\ ] ] for regions i and ii ( @xmath277 ) , and @xmath278 for regions iii and iv ( @xmath279 ) . the solutions [ eqs.([eq146])([eq149 ] ) ] are illustrated in fig .
[ fig7 ] . using eq .
( [ eq52c ] ) we obtain the asymptotic solution as @xmath187 @xmath280 the exact solution given by eq .
( [ eq148 ] ) and asymptotic solution given by eq .
( [ eq149a ] ) are compared in fig.[fig8 ] ( line b ) for @xmath281 .
the next example we consider corresponds to the initial linear density profile @xmath282 here we repeat the intermediate steps in the previous example . for a cold beam , @xmath252 . substituting eq.([eq93a ] ) into eq .
( [ eq74e ] ) and integrating , we obtain @xmath283^{3/2}\theta(z < c_0).\label{eq94}\ ] ] next we substitute eq .
( [ eq94 ] ) into the integral @xmath284^{3/2},\label{eq95}\end{aligned}\ ] ] and define @xmath285 where @xmath120 and @xmath286 are introduced in eq .
( [ eq79 ] ) . by introducing the new integration variable @xmath287 defined by @xmath288 the integral in eq .
( [ eq95 ] ) can be expressed as @xmath289^{3/2}\over ( a+b)^3}\int^{\pi/2}_0 d\alpha{\cos^4(\alpha)[1-k^2p^2\sin^2(\alpha)]^{3/2}\over [ 1-k^2\sin^2(\alpha)]^4},\label{eq98}\end{aligned}\ ] ] where @xmath290 and @xmath291 .
the integral in eq .
( [ eq98 ] ) can be expressed in terms of elliptic integrals . finally ,
using the definitions @xmath292 and @xmath293 , and using eqs.([eq79 ] ) , ( [ eq82 ] ) , and ( [ eq83 ] ) , we obtain after some lengthy algebra the solution in regions i and iii , @xmath294=2\bar v , \nonumber \\ \bar x&=&{4\over \pi}{(1 + 2\bar v^2-\bar c^2 ) \bar
c\over \sqrt{(1+\bar c)^2-\bar v^2}}\bigg[\pi\left({1+\bar v -\bar c \over1+\bar v+\bar c } , { ( 1-\bar c)^2-\bar v^2\over(1+\bar c)^2-\bar v^2 } \right)\label{eq100}\\ & + & \pi\left({1-\bar v -\bar c \over1-\bar v+\bar c } , { ( 1-\bar c)^2-\bar v^2\over(1+\bar c)^2-\bar v^2 } \right)-k\left({(1-\bar c)^2-\bar v^2\over(1+\bar c)^2-\bar v^2 }
\right)\bigg]=1 + 2\bar v^2-\bar c^2 , \nonumber\end{aligned}\ ] ] and in regions ii and iv , @xmath295,\nonumber\\ \bar x&=&{2\over 3\pi}{1\over \bar c^2 \sqrt{(1+\bar c)^2-\bar v^2}}\bigg[6\bar c^3(1-\bar c^2 + 2\bar v^2 ) \bigg\{\pi\left({1+\bar v -\bar c \over1+\bar v+\bar c } , { ( 1-\bar c)^2-\bar v^2\over(1+\bar c)^2-\bar v^2 } \right)\label{eq102}\\ & -&\pi\left({1-\bar v -\bar c \over1-\bar v+\bar c } , { ( 1-\bar c)^2-\bar v^2\over(1+\bar c)^2-\bar v^2 } \right)\bigg\}-2\bar c\bar v(4 - 4\bar v^2+\bar c(2+\bar c-3\bar c^2 + 6\bar v^2))k\left({(1-\bar c)^2-\bar v^2\over(1+\bar c)^2-\bar v^2 } \right)\nonumber \\ & + & \bar v((1+\bar c)^2-\bar v^2)(4+\bar c^2 - 4\bar v^2)e\left({(1-\bar c)^2-\bar v^2\over(1+\bar c)^2-\bar v^2 } \right)\bigg],\nonumber\end{aligned}\ ] ] where @xmath264 , @xmath265 , @xmath296 and @xmath297 . here ,
@xmath243 , @xmath298 and @xmath299 are the complete elliptic integrals of the first , second , and the third kinds , respectively @xcite .
the solution to the expansion problem for the linear density profile eq .
( [ eq93a ] ) is given by eqs.([eq99])([eq102 ] ) and is illustrated in fig.[fig9 ] .
using eqs .
( [ eq99 ] ) and ( [ eq100 ] ) , we can rewrite the solution in regions i and iii as @xmath300 this simple flow with uniform acceleration in regions i and iii exists only until @xmath301 . at this time the characteristic q overtakes the edge of the beam , and
the flow for @xmath302 is entirely in regions ii and iv and is given by eqs.([eq101 ] ) and ( [ eq102 ] ) .
using eqs .
( [ eq74d ] ) and ( [ eq94 ] ) we obtain the asymptotic solution as @xmath303 @xmath304^{3/2},\ \ \ \ v(x , t)={x\over t},\ \ \text{for}\ \ t\rightarrow \infty . \label{eq104a}\end{aligned}\ ] ] the exact solution given by eqs .
( [ eq101 ] ) and ( [ eq102 ] ) and asymptotic solution given by eq .
( [ eq104a ] ) are compared in fig .
[ fig6 ] ( line a ) for @xmath270 . for a pressure - dominated beam , @xmath271 .
hence , for the linear density profile in eq .
( [ eq93a ] , the initial profile for @xmath138 is @xmath305\theta(c < c_0),\label{eq136}\ ] ] where @xmath212 is the initial beam half - length , and @xmath306 is the sound speed at @xmath148 .
substituting eq .
( [ eq136 ] ) into eqs.([eq49])([eq52 ] ) , we obtain the solution in region i , @xmath307 in region ii , @xmath308 and in region iv , @xmath309 region iii has disappeared . indeed , using eqs .
( [ eq51 ] ) we obtain @xmath310 and @xmath311 , so that the entire region consist of just one point .
the corresponding solutions [ eqs.([eq137])([eq139 ] ) ] are illustrated in fig.[fig10 ] .
using eqs .
( [ eq52c ] ) and ( [ eq136 ] ) we obtain the asymptotic solution as @xmath187 @xmath312 the exact solution given by eq .
( [ eq139 ] ) and asymptotic solution given by eq .
( [ eq139a ] ) are compared in fig.[fig8 ] ( line a ) for @xmath281 . as discussed in sec ii , the general flow consists of regions where either a simple wave solution [ eq .
( [ eq28 ] ) ] or a general solution [ solution of eq .
( [ eq13 ] ) together with eq.([eq7 ] ) ] is applicable . up to now we have considered flows with no simple wave regions .
this is guaranteed provided the initial density profile is smooth everywhere except at the beam edge where the sound speed is zero , @xmath313 . here , we consider an example with an initial flat - top density profile , and a corresponding discontinuity in density at the beam edge , i.e. , @xmath314 this problem is equivalent to the one - dimensional expansion of a uniform - density gas into vacuum in a vessel in which the end walls are instantly removed .
the @xmath47 plane is shown in fig.[fig11 ] .
the flow consists of three regions . in region
i the gas is at rest , and the information that the walls have been removed , which is carried by the @xmath87 characteristic p into the gas , does not reach this region .
region ii is the region occupied by a simple rarefaction wave which is centered at @xmath137 and @xmath315 in the @xmath47 plane , and is described by the simple wave solution in eq .
( [ eq28 ] ) .
region iii is the region of interference of this wave and and its reflection from the origin ( or another rarefaction wave coming from the other end of the gas region ) .
this region is described by the general solution in eq.([eq13 ] ) together with eq .
( [ eq7 ] ) .
regions ii and iii are separated by the @xmath86 characteristic q. on this characteristic the boundary condition in eq .
( [ eq31 ] ) holds . to determine the function @xmath74 in eqs .
( [ eq28 ] ) and ( [ eq31 ] ) , we note that for @xmath137 , @xmath315 and therefore @xmath316 .
also note that the characteristics @xmath86 bring the value of the rieman invariant @xmath317 to all points of region ii .
therefore , the solution in region ii is given by @xmath318 where @xmath101 , and the boundary condition for @xmath44 on the separating characteristic q where @xmath143 is given by @xmath319 the second boundary condition [ eq . ( [ eq41 ] ) ] is given by @xmath320 the function @xmath44 satisfying the boundary condition in eq.([eq120 ] ) is given by @xmath321.\label{eq121}\ ] ] using eqs .
( [ eq119 ] ) and ( [ eq121 ] ) , we obtain the integral equation for the function @xmath228 , @xmath322+f[(c_0-v/2)t+v/2]\}.\label{eq122}\ ] ] by setting @xmath323 in eq .
( [ eq122 ] ) we note that the function @xmath228 has the form @xmath324 . substituting this expression for @xmath325 into eq .
( [ eq122 ] ) , and changing the integration variable to @xmath326 , we obtain the integral equation for the function @xmath3 , @xmath327 ) } } = vx_0.\label{eq123}\ ] ] changing the integration variables in eq .
( [ eq123 ] ) according to @xmath328 , and introducing the new function @xmath329 $ ] , we obtain an integral equation of the abel type , @xmath330 where @xmath331 .
equation ( [ eq124 ] ) is easily solved using the abel transform described in appendix a. we obtain @xmath332 finally , since @xmath333 , we obtain @xmath334 next we substitute eq .
( [ eq126 ] ) into the integral @xmath335^{1/2},\label{eq127}\end{aligned}\ ] ] and define @xmath336 where @xmath120 and @xmath286 are introduced in eqs .
( [ eq79 ] ) . by introducing the new integration variable @xmath287 defined by @xmath337 the integral in eq .
( [ eq127 ] ) can be rewritten as @xmath338^{1/2}\over ( a+b)}\int^{\pi/2}_0 d\alpha{\cos^2(\alpha)[1-k^2p^2\sin^2(\alpha)]^{1/2}\over [ 1-k^2\sin^2(\alpha)]^2},\label{eq130}\end{aligned}\ ] ] where @xmath290 and @xmath291 .
the integral in eq .
( [ eq130 ] ) can be expressed in terms of elliptic integrals . finally , using the definition @xmath339 , and making use of eqs .
( [ eq79 ] ) , ( [ eq82 ] ) and ( [ eq83 ] ) , we obtain after some lengthy algebra the solution in region iii , @xmath340e\left[{(1-\bar c)^2-\bar v^2\over ( 1+\bar c)^2-\bar v^2}\right]-2\bar c
k\left[{(1-\bar c)^2-\bar v^2\over ( 1+\bar c)^2-\bar v^2}\right]\right\},\nonumber\\ \label{eq131}\\ & & \bar x={4\over \pi \bar c^2\sqrt{(1+\bar c)^2-\bar v^2}}\bigg\{\bar v[(1+\bar c)^2)-\bar v^2]e\left[{(1-\bar c)^2-\bar v^2\over ( 1+\bar c)^2-\bar v^2}\right]-\bar v\bar c(2+\bar c ) k\left[{(1-\bar c)^2-\bar v^2\over ( 1+\bar c)^2-\bar v^2}\right]\nonumber \\ & & + \pi\left[{1-\bar c-\bar v\over 1+\bar c-\bar v},{(1-\bar c)^2-\bar v^2\over ( 1+\bar c)^2-\bar v^2}\right]-\pi\left[{1-\bar c+\bar v\over 1+\bar c+\bar v},{(1-\bar c)^2-\bar v^2\over ( 1+\bar c)^2-\bar v^2}\right]\bigg\},\label{eq132}\end{aligned}\ ] ] where @xmath264 , @xmath265 , @xmath296 and @xmath297 . here ,
@xmath243 , @xmath298 and @xmath299 are the complete elliptic integrals of the first , second , and the third kinds , respectively @xcite .
the solution in region ii is given by eqs.([eq117 ] ) and ( [ eq118 ] ) with @xmath341 , i.e. , @xmath342 using eqs .
( [ eq133 ] ) and ( [ eq134 ] ) we can determine the trajectory of the beam edge and the characteristics q and p. at the beam edge , @xmath343 , and therefore from eq .
( [ eq133 ] ) we obtain @xmath344 . on the characteristic p , @xmath323 , and from eq .
( [ eq134 ] ) we obtain @xmath345 . on the characteristic q , @xmath346 .
integrating this equation , we obtain @xmath347 the solutions given by eqs.([eq131])([eq134 ] ) are illustrated in fig .
[ fig12 ] . using eqs .
( [ eq74d ] ) and ( [ eq126 ] ) we obtain the asymptotic solution in region iii @xmath348 $ ] as @xmath303 @xmath349 the asymptotic solution in region ii @xmath350 $ ] is still given by eqs .
( [ eq133 ] ) and ( [ eq134 ] ) . the exact solution given by eqs .
( [ eq131])([eq134 ] ) and asymptotic solution given by eqs .
( [ eq135a ] ) and ( [ eq133 ] ) are compared in fig .
[ fig6 ] ( line c ) for @xmath270 .
the function @xmath44 satisfying the boundary condition in eq.([eq120 ] ) is given by @xmath351 using the boundary condition on the characteristic q , it follows that @xmath352 for @xmath353 , and we obtain @xmath354 substituting @xmath323 into eq .
( [ eq151 ] ) , we find that @xmath175 and therefore @xmath355 , or @xmath356 .
using eq .
( [ eq150 ] ) , we obtain @xmath357 and using eq .
( [ eq7 ] ) , we obtain @xmath358 and @xmath359 . therefore , the solution in region iii is given by @xmath360 the solution in region ii is given by eqs.([eq117 ] ) and ( [ eq118 ] ) with @xmath361 , i.e. , @xmath362 in region i , the gas is undisturbed : @xmath363 and @xmath131 , for @xmath364 and @xmath365 .
the characteristic q is given by the equation @xmath366 , and the beam edge is given by @xmath367 .
the solutions given by eqs.([eq153 ] ) and ( [ eq154 ] ) are illustrated in fig.[fig13 ] . since @xmath44 in eq .
( [ eq152 ] ) is a linear function of @xmath89 and is independent of @xmath50 , the asymptotic @xmath368 solution coincides with the exact solution in eq .
( [ eq153 ] ) . up to this point
we have considered flows which do not form shocks and therefore are time - reversible . for such flows ,
the compression problem is equivalent to the time - reversed expansion problem . in this section ,
we consider an example of fluid flow which forms shocks . for simplicity
, we consider here a beam without sharp edges in which @xmath133 decreases to zero monotonically as @xmath369 . in particular , we consider the initial density profile given by @xmath370 expanding flows with initially smooth profiles extending to @xmath369 such as in eq .
( [ eq105 ] ) are entirely in regions i and ii .
indeed regions i and ii are separated from regions iii and iv by the @xmath87 characteristic p with @xmath371 ( @xmath101 ) . however , for profiles such as eq.([eq105 ] ) , @xmath372 for all @xmath373 , and therefore the entire region @xmath373 maps into regions i and ii in the @xmath152 plane ( see fig .
[ fig3 ] ) . as a result
, the solution for the flow is given by eqs .
( [ eq49 ] ) and ( [ eq50 ] ) for pressure - dominated beams , and by eqs .
( [ eq70c ] ) and ( [ eq70e ] ) , with @xmath230 defined in eq .
( [ eq70 ] ) for cold beams . for a cold beam , @xmath252 , and therefore @xmath374 .
substituting eq .
( [ eq105 ] ) into eq.([eq74e ] ) and integrating , we obtain @xmath375 substituting eq .
( [ eq107 ] ) into the integral for @xmath376 and integrating , we obtain @xmath377,\label{eq108}\end{aligned}\ ] ] and @xmath378 where @xmath120 and @xmath286 are introduced in eq .
( [ eq79 ] ) . using the definitions @xmath379 and @xmath380 , and making use of eqs.([eq79 ] ) , ( [ eq82 ] ) and ( [ eq83 ] )
, we obtain the solution in region i , @xmath381,\label{eq111}\end{aligned}\ ] ] and in region ii , @xmath382.\label{eq113}\end{aligned}\ ] ] equations ( [ eq110])([eq113 ] ) can be partially inverted to give @xmath383 where @xmath384 and @xmath385 .
the solutions given by eqs .
( [ eq114 ] ) and ( [ eq115 ] ) are illustrated in fig.[fig14 ] . as evident from eq .
( [ eq114 ] ) , @xmath386 in some regions , which means that the regions with higher density accelerated faster than the regions with lower density , and eventually multi - valued flow is formed ( see fig . [ fig14 ] ) .
this is unlike the previous examples where @xmath387 for all @xmath72 , and there was no multi - valued flow . for a pressure - dominated beam , @xmath271 . using eq .
( [ eq105 ] ) and eqs .
( [ eq62 ] ) and ( [ eq63 ] ) , we obtain the implicit solution in regions i and ii , @xmath388}+ { 1\over \cosh^2[\bar x-(\bar v-\bar c)\bar t]}\right\},\label{eq140}\\ & & \bar v(x , t)={1\over 2}\left\{{1\over \cosh^2[\bar x-(\bar v+\bar c)\bar t]}- { 1\over \cosh^2[\bar x-(\bar v-\bar c)\bar t]}\right\}.\label{eq141}\end{aligned}\ ] ] the solutions given by eqs .
( [ eq140 ] ) and ( [ eq141 ] ) are illustrated in fig .
[ fig15 ] .
in this section we consider the beam shaping problem referred to in sec .
i. that is , given an initial line density profile @xmath12 at time @xmath137 and final line density profile @xmath389 at time @xmath390 , what are the initial and finial velocity profiles , @xmath391 and @xmath392 respectively .
the beam shaping stage is necessary to prepare the beam density profile for the final drift compression discussed in previous sections . here , as in previous sections , we analyze the time - reversed problem .
therefore , the initial density profile @xmath12 for the time - reversed problem is given by eqs.([eq74d ] ) and ( [ eq74e ] ) for a cold beam , or by eq.([eq52c ] ) for a pressure - dominated beam , and the final density profile @xmath10 illustrated schematically in fig.[fig1 ] . during the beam shaping stage ,
the longitudinal pressure and electric field are negligible and the beam dynamics is governed by free convection decribed by @xmath393 here , eq .
( [ sh2 ] ) follows from multiplying eq .
( [ eq11 ] ) by @xmath394 , and is equivalent to eq .
( [ eq1 ] ) .
equation ( [ sh1 ] ) implies that the function @xmath50 is constant along the characteristic given by @xmath395 , which therefore correspond to straight lines given by @xmath396,\label{sh3}\ ] ] where @xmath397 and @xmath398=v$ ] .
equation ( [ sh3 ] ) gives a general solution to eq .
( [ sh1 ] ) for the velocity profile @xmath25 .
substituting eq .
( [ sh3 ] ) into eq .
( [ sh2 ] ) , we obtain @xmath399 , where @xmath400 is an arbitrary function of @xmath50 . using the initial condition that at @xmath137 , @xmath401
, we obtain the solution to eqs.([sh2 ] ) , @xmath402.$ ] note that @xmath403=v'(x - vt)$ ] , and @xmath404\equiv \lambda(x - vt)$ ] , where @xmath405 is the initial density profile as a function of @xmath77 .
therefore , the solution to eqs .
( [ sh1 ] ) and ( [ sh2 ] ) is given by @xcite @xmath406,\label{sh5}\\ & & \lambda(x , t)={\lambda[x - v(x , t)t]\over 1+t v'[x - v(x , t)t]}.\label{sh6}\end{aligned}\ ] ] setting @xmath390 and introducing new the function @xmath407 , we can rewrite eq .
( [ sh5 ] ) as @xmath408=\left\{u_{in}[x - v_f(x)t]-[x - v_f(x)t_{shape}]\right\}/t_{shape}$ ] , or equivalently , @xmath409 $ ] . finally ,
using the definition of the function @xmath410 , we can rewrite eqs.([sh5 ] ) and ( [ sh6 ] ) in a compact and manifestly time - reversible form giving the finial formal solution to the beam shaping problem , i.e. , @xmath411 here , @xmath412 is the inverse of function @xmath410 such that @xmath413\equiv x$ ] , and we have assumed that @xmath414 and @xmath415 .
examples applying results in eqs .
( [ sh7 ] ) and ( [ sh8 ] ) can be found in ref .
to summarize , we have studied the longitudinal drift compression of an intense charged particle beam using a one - dimensional warm - fluid model .
we have reformulated the drift compression problem as the time - reversed expansion problem of the beam with arbitrary line density profile and zero velocity profile .
we have obtained exact analytical solutions to the expansion problem for the two important cases corresponding to a cold beam , and a pressure - dominated beam , using a general formalism which reduces the system of warm - fluid equations to a linear second - order partial differential equation .
we obtained simple approximate analytical formulas connecting the initial and final line density profile and flow velocity profile for these two cases .
the asymptotic velocity profiles are linear in both cases , and correspond to free expansion as @xmath416 . the scaled density profile for a pressure - dominated beam far from the compression point
was shown to be the functional inverse of the compressed density profile in eq .
( [ eq52c ] ) . for a cold beam ,
the profiles are connected by the abel transform [ eqs.([eq74d ] ) and ( [ eq74e ] ) ] .
the general solution has been illustrated for parabolic , linear , and flat - top initial ( compressed ) line density profiles .
for the case of a parabolic density profile , we have recovered the familiar self - similar solution @xcite .
we have illustrated the formation of multi - valued flow with the exactly - solvable example in eq .
( [ eq105 ] ) , and identified the conditions for shock - free compression .
this research was supported by the u.s .
department of energy .
it is a pleasure to acknowledge the benefit of useful discussions with igor kaganovich and hong qin .
here we use the following definition of the abel transform @xmath417=2\int_z^{\infty}{g(x)xdx\over \sqrt{x^2-z^2}}.\label{app1}\ ] ] the inverse abel transform is given by @xmath418=-{1\over \pi}\int_x^{\infty}{df(z)\over dz}{dz\over \sqrt{z^2-x^2}}.\label{app2}\ ] ] the fact that eq .
( [ app2 ] ) is indeed the inverse of the abel transform in eq .
( [ app1 ] ) can be checked by direct substitution of eq .
( [ app2 ] ) into eq .
( [ app1 ] ) .
changing the order of integration leads to @xmath419 making the change of variables , @xmath420 , the final integral in eq.([app3 ] ) is evaluated to be @xmath421 and eq .
( [ app3 ] ) becomes @xmath422 therefore , for functions @xmath230 such that @xmath423 , eq.([app5 ] ) is an identity .
99 r. c. davidson and h. qin , _ physics of intense charged particle beams in high energy accelerators _
( world scientific , singapore , 2001 ) , and references therein .
m. reiser , _ theory and design of charged particle beams _ ( wiley , new york , 1994 ) .
j. d. lawson , _ the physics of charged - particle beams _ ( oxford science publications , new york , 1988 ) .
a. w. chao , _ physics of collective beam instabilities in high energy accelerators _ ( wiley , new york , 1993 ) .
d. a. edwards and m. j. syphers , _ an introduction to the physics of high - energy accelerators _
( wiley , new york , 1993 ) . i. m. kapchinskij and v. v. vladimirskij , in _ proceedings of the international conference on high enegy accelerators and instrumentation _ ( cern scientific information service , geneva , 1959 ) , p. 274 .
r. l. gluckstern , in _ proceedings of the 1970 proton linear accelerator conference , batavia , il _ , edited by m. r. tracy ( national accelerator laboratory , batavia , il , 1971 ) , p. 811 .
wang and i. smith , part .
* 12 * , 247 ( 1982 ) .
j. hofmann , l. j. laslett , l. smith , and i. haber , part .
* 13 * , 145 ( 1983 ) .
j. struckmeier , j. klabunde , and m. reiser , part .
accel . * 15 * , 47 ( 1984 ) .
i. hofmann and j. struckmeier , part .
* 21 * , 69 ( 1987 ) .
j. struckmeier and i. hofmann , part .
accel . * 39 * , 219 ( 1992 ) . n. brown and m. reiser , phys .
plasmas * 2 * , 965 ( 1995 ) .
r. c. davidson and h. qin , phys . rev .
beams * 2 * , 114401 ( 1999 ) .
r. l. gluckstern , w. -h .
cheng , and h. ye , phys .
lett . * 75 * , 2835 ( 1995 ) .
r. c. davidson and c. chen , part .
* 59 * , 175 ( 1998 ) . c. chen , r. pakter , and r. c. davidson , phys .
. lett . * 79 * , 225 ( 1997 ) . c. chen and r. c. davidson , phys .
e * 49 * , 5679 ( 1994 ) .
r. c. davidson , w. w. lee , and p. stoltz , phys .
plasmas * 5 * , 279 ( 1998 ) .
r. c. davidson , phys .
lett . * 81 * , 991 ( 1998 ) .
r. c. davidson , phys .
plasmas * 5 * , 3459 ( 1998 ) .
p. stoltz , r. c. davidson , and w. w. lee , phys .
plasmas * 6 * , 298 ( 1999 ) .
w. w. lee , q. qian , and r. c. davidson , phys .
a * 230 * , 347 ( 1997 ) .
q. qian , w. w. lee , and r. c. davidson , phys .
plasmas * 4 * , 1915 ( 1997 ) . s. i. tzenov and r. c. davidson , phys . rev .
beams * 5 * , 021001 ( 2002 ) .
r. c. davidson and h. qin , phys . rev .
beams * 4 * , 104401 ( 2001 ) .
r. c. davidson , h. qin , and p. j. channell , phys . rev .
beams * 2 * , 074401 ( 1999 ) ; * 3 * , 029901 ( 2000 ) .
r. c. davidson , a. friedman , c. m. celata , d. r. welch , et al . ,
laser and particle beams * 20 * , 377 ( 2002 ) .
e. g. harris , phys .
* 2 * , 34 ( 1959 ) .
e. a. startsev , r. c. davidson and h. qin , physics of plasmas * 9 * , 3138 ( 2002 ) .
e. a. startsev , r. c. davidson and h. qin , laser and particle beams * 20 * , 585 ( 2002 ) .
e. a. startsev , r. c. davidson and h. qin , physical review special topics on accelerators and beams * 6 * , 084401 ( 2003 ) .
e. a. startsev and r. c. davidson , physical review special topics on accelerators and beams * 6 * , 044401 ( 2003 ) .
r. a. kishek , p. g. oshea , and m. reiser , phys .
lett . * 85 * , 4514 ( 2000 ) .
j. haber , a. friedman , d. p. grote , s. m. lund , and r. a. kishek , phys .
plasmas * 6 * , 2254 ( 1999 ) . a. friedman , d. p. grote , and i. haber , phys
. fluids b * 4 * , 2203 ( 1992 ) .
e. s. weibel , phys . rev
* 2 * , 83 ( 1959 ) .
e. a. startsev and r. c. davidson , physics of plasmas * 10 * , 4829 ( 2003 ) .
r. c. davidson , d. a. hammer , i. haber and c. e. wagner , phys .
fluids * 15 * , 317 ( 1972 ) .
r. lee and m. lampe , phys .
lett . * 31 * , 1390 ( 1973 ) . c. a. kapetanakos , appl .
. lett . * 25 * , 484 ( 1974 ) .
m. honda , j. meyer - ter - vehn , a. pukhov , phys .
lett . * 85 * , 2128 ( 2000 ) .
v. k. neil and a. m. sessler , rev .
instr . * 36 * , 429 ( 1965 ) .
e. p. lee , in _ proceedings of the 1981 linear accelerator conference _ , los alamos national laboratory report la-9234-c , pp . 263265 .
e. p. lee , particle accelerators * 37 * , 307 ( 1992 ) .
r. c. davidson , h. qin and g. shvets , physical review special topics on accelerators and beams * 6 * , 104402 ( 2003 ) .
d. neuffer , b. colton , d. fitzgerald , t. hardek , r. hutson , r. macek , m. plum , h. thiessen , and t. s. wang , nucl .
methods phys .
a321 * , 1 ( 1992 ) .
r. j. macek , et al .
, _ proceedings of the 2001 particle accelerator conference _ * 1 * , 688 ( 2001 ) .
r. c. davidson , h. qin , and t. -s .
wang , physics letters * a252 * , 213 ( 1999 ) .
r. c. davidson , h. qin , p. h. stoltz , and t. -s .
wang , physical review special topics on accelerators and beams * 2 * , 054401 ( 1999 ) .
r. c. davidson and h. qin , physics letters * a270 * , 177 ( 2000 ) . h. qin , r. c. davidson and w. w. lee , physical review special topics on accelerators and beams * 3 * , 084401 ( 2000 ) ; 3 , 109901 ( 2000 ) .
r. c. davidson and h. s. uhm , physics letters * a285 * , 88 ( 2001 ) .
h. qin , r. c. davidson , e. a. startsev and w. w. lee , laser and particle beams * 21 * , 21 ( 2003 )
wang , p. j. channell , r. j. macek , and r. c. davidson , physical review special topics on accelerators and beams * 6 * , 014204 ( 2003 ) .
h. qin , e. a. startsev and r. c. davidson , physical review special topics on accelerators and beams * 6 * , 014401 ( 2003 ) .
h. qin , physics of plasmas * 10 * , 2708 ( 2003 ) .
e. p. lee , phys .
fluids * 21 * , 1327 ( 1978 ) .
e. j. lauer , r. j. briggs , t. j. fesendon , r. e. hester , and e. p. lee , phys .
fulids * 21 * , 1344 ( 1978 ) . m. n. rosenbluth , phys .
fluids * 3 * , 932 ( 1960 ) .
h. s. uhm and m. lampe , phys .
fluids * 23 * , 1574 ( 1980 ) . h. s. uhm and m. lampe , phys .
fluids * 24 * , 1553 ( 1981 ) .
r. f. fernsler , s. p. slinker , m. lampe , and r. f. hubbard , phys .
plasmas * 2 * , 4338 ( 1995 ) , and references therein . h. s. uhm and r. c. davidson , physical review special topics on accelerators and beams * 6 * , 034204 ( 2003 ) . h.
s. uhm , r. c. davidson and i. d. kaganovich , physics of plasmas * 8 * , 4637 ( 2001 ) .
g. joyce and m. lampe , phys .
fluids * 26 * , 3377 ( 1983 ) . h. s. uhm and r. c. davidson , phys .
fluids * 23 * , 1586 ( 1980 ) .
i. d. kaganovich , g. shvets , e. startsev and r. c. davidson , physics of plasmas * 8 * , 4180 ( 2001 ) .
i. kaganovich , e. a. startsev and r. c. davidson , laser and particle beams * 20 * , 497 ( 2002 ) .
d. v. rose , p. f. ottinger , d. r. welch , b. v. oliver and c. l. olson , phys .
plasmas * 6 * , 4094 ( 1999 ) .
d. r. welch , d. v. rose , b. v. oliver , t. c. genoni , c. l. olson and s. s. yu , phys .
plasmas * 9 * , 2344 ( 2002 ) .
i. hofmann , z. naturforsch .
* 37a * , 939 ( 1982 ) . i. hofmann , laser and particle beams * 3 * , 1 ( 1985 ) . o. boine - frankenheim , i. hofmann and g. rumolo , phys .
lett . * 82 * , 3256 ( 1999 ) .
l. k. spentzouris , j. -f .
ostiguy and p. l. colestock , phys .
lett . * 76 * , 620 ( 1996 ) . l. k. spentzouris , p. l. colestock and c. bhat , proceedings of the 1999 particle accelerator conference * 1 * , 114 ( 1999 ) .
r. fedele , g. miele , l. palumbo and v. g. vaccaro , phys .
* a179 * , 407 ( 1993 ) . h. schamel , phys .
lett . * 79 * , 2811 ( 1997 ) .
h. schamel and r. fedele , phys .
plasmas * 7 * , 3421 ( 2000 ) .
a. hofmann , cern report no . 77 - 13 ( 1977 ) .
d. neuffer , particle accelerators , * 11 * , 23 , 1980 . c. k. allen , n. brown and m. reiser , particle accelerators * 45 * , 149 ( 1994 )
. w. m. sharp , a. friedman and d. p. grote , fusion engineering and design * 32 * , 201 ( 1996 ) .
j. g. wang , h. suk , d. x. wang and m. reiser , phys .
72 * , 2029 ( 1994 ) .
r. c. davidson and e. a. startsev , physical review special topics on accelerators and beams * 7 * , 024401 ( 2004 ) .
r. c. davidson and s. strasburg , physics of plasmas * 7 * , 2657 ( 2000 ) . s. m. lund and r. c. davidson , physics of plasmas * 5 * , 3028 ( 1998 ) .
r. c. davidson , h. qin , s. i. tzenov , and e. a. startsev , physical review special topics on accelerators and beams 5 , 084402 ( 2002 ) .
h. qin , r. c. davidson , j. j. barnard and e. p. lee , in _ proceedings of the 2003 particle accelerator conference _
, 2658 ( 2003 ) . h. qin and r. c. davidson , physical review special topics on accelerators and beams * 5 * , 03441 ( 2002 ) .
h. qin and r. c. davidson , laser and particle beams * 20 * , 565 ( 2002 ) .
i. haber , in _ proccedings of the symposium on accelerator aspects of heavy ion fusion , darmstadt , 1982 _
( gsi , darmstadt , 1982 ) , p. 372 . i. hofmann and i. bozsik , in _ proccedings of the symposium on accelerator aspects of heavy ion fusion , darmstadt , 1982 _
( gsi , darmstadt , 1982 ) , p. 362 .
j. bisognano , e. p. lee and j. w .- k .
mark , llnl report , no .
3 - 28 , 1985 .
d. d .- m .
ho , s. t. brandon and e. p. lee , part .
* 35 * , 15 ( 1991 ) .
m. j. l. de hoon , phd .
thesis , university of california , berkeley , 2001 .
e. p. lee and j. j. barnard , in _ proceedings of the 2001 particle accelerator conference , chicago , 2001 _ , p. 2928 .
h. qin , c. jun , r. c. davidson and p. heitzenroeder , in _ proceedings of the 2001 particle accelerator conference , chicago , 2001 _ , p. 1761
* figure captions * fig.1 : schematic of the two stages of drift compression described in sec .
i , the _ beam shaping _ stage and the _ drift compression _ stage .
fig.2 : the area in the @xmath47 plane occupied by the four regions of flow .
fig.3 : the area in the @xmath152 plane occupied by the four regions of flow .
fig.4 : schematic in @xmath424 phase space of the flow for a pressure - dominated beam .
fig.5 : plots of the normalized density @xmath425 ( a ) , and the flow velocity @xmath426 ( b ) , at @xmath427 as functions of @xmath77 for a cold beam .
the initial density profile ( dotted line ) is given by eq .
( [ eq75a ] ) .
fig.6 : plots of the normalized density @xmath425 as a function of @xmath77 for a cold beam at @xmath270 .
the solid lines are the exact solutions .
the dotted lines are the approximate solutions given by eq .
( [ eq74d ] ) .
the initial profiles are given by : ( a ) eq .
( [ eq93a ] ) , ( b ) eq . ( [ eq75a ] ) , and ( c ) eq . ( [ eq116 ] ) .
fig.7 : plots of the normalized density @xmath425 for @xmath428 ( a ) and for @xmath429 ( c ) , and the flow velocity @xmath426 for @xmath428 ( b ) and for @xmath429 ( d ) , as functions of @xmath77 for a cold beam .
the initial density profile ( dotted line ) is given by eq .
( [ eq93a ] ) .
fig.8 : plots of the normalized density @xmath425 as a function of @xmath77 for a pressure - dominated beam at @xmath430 .
the solid lines are the exact solutions .
the dotted lines are the approximate solutions given by eq .
( [ eq52c ] ) .
the initial profiles are given by : ( a ) eq .
( [ eq93a ] ) , ( b ) eq . ( [ eq75a ] ) , and ( c ) eq . ( [ eq116 ] ) .
fig.9 : plots of the normalized density @xmath425 for @xmath428 ( a ) and for @xmath429 ( c ) , and the flow velocity @xmath431 for @xmath428 ( b ) and for @xmath429 ( d ) , as functions of @xmath77 for a pressure - dominated beam . the initial density profile ( dotted line ) is given by eq.([eq75a ] ) .
fig.10 : plots of the normalized density @xmath425 for @xmath428 ( a ) and for @xmath429 ( c ) , and the flow velocity @xmath431 for @xmath428 ( b ) and for @xmath429 ( d ) , as functions of @xmath77 for a pressure - dominated beam .
the initial density profile ( dotted line ) is given by eq.([eq93a ] ) .
fig.11 : the area in the @xmath47 plane occupied by the three regions of flow .
fig.12 : plots of the normalized density @xmath425 for @xmath428 ( a ) and for @xmath432 ( c ) , and the flow velocity @xmath426 for @xmath428 ( b ) and for @xmath432 ( d ) , as functions of @xmath77 for a cold beam .
the initial density profile ( dotted line ) is given by eq .
( [ eq116 ] ) .
fig.13 : plots of the normalized density @xmath425 for @xmath428 ( a ) and for @xmath433 ( c ) , and the flow velocity @xmath431 for @xmath428 ( b ) and for @xmath429 ( d ) , as functions of @xmath77 for a pressure - dominated beam .
the initial density profile ( dotted line ) is given by eq .
( [ eq116 ] ) .
fig.14 : plots of the normalized density @xmath425 ( a ) and flow velocity @xmath426 ( b ) at @xmath434 as functions of @xmath77 for a cold beam .
the initial density profile ( dotted line ) is given by eq .
( [ eq105 ] ) .
fig.15 : plots of the normalized density @xmath425 ( a ) and flow velocity @xmath431 ( b ) at @xmath435 as functions of @xmath77 for a pressure - dominated beam .
the initial density profile ( dotted line ) is given by eq.([eq105 ] ) . | to achieve high focal spot intensities in heavy ion fusion , the ion beam must be compressed longitudinally by factors of ten to one hundred before it is focused onto the target .
the longitudinal compression is achieved by imposing an initial velocity profile tilt on the drifting beam . in this paper ,
the problem of longitudinal drift compression of intense charged particle beams is solved analytically for the two important cases corresponding to a cold beam , and a pressure - dominated beam , using a one - dimensional warm - fluid model describing the longitudinal beam dynamics . |
to obtain the dark matter distribution at large scale around galaxies , much farther than the extent of rotation curves , astronomers have since a long time used the kinematics of satellite galaxies ( erickson et al 1987 , zaritsky et al 1993 , 1994 , 1997 ) . because the number of satellites per galaxy is very small , the technique consists in stacking the data on many galaxy - satellite pairs , in common bins of host luminosity . in the 1990s ,
the number of satellites were counted in hundreds , and the statistics were not sufficient to provide significant results .
for instance with 115 satellites at distances lower than 500kpc , the probability to find nsat satellites was fitted to p(nsat ) = 0.4@xmath1 ( zaritsky et al 1997 ) .
no dependence of velocity dispersion of the satellites with radius , nor with the host mass , was found .
the advent of rich surveys such as the sdss and 2df grs have given more statistical value to the satellite studies , including now nsat = a few thousands ( mckay et al 2002 , brainerd & specian 2003 , prada et al 2003 , van den bosch et al 2004 ) .
while some studies still find a velocity dispersion @xmath2 flat with radius , most of them now see the decrease of @xmath2 with radius , and the increase with the host luminosity or mass . however , there are still contradictory results , as reviewed by norberg et al ( 2008 ) .
they found that the velocity dispersion within a projected radius of 175@xmath3kpc is increasing as the square root of luminosity for ellipticals , implying a constant m / l ratio with mass .
for spirals the slope is higher , @xmath4 and the mass - to - light ratio is increasing with luminosity .
m / l is 3 - 10 times higher for ellipticals than for spirals , at the same luminosity l@xmath5 . for the milky way luminosity bin , m(175@xmath3kpc)= @xmath6 , but there is a large scatter , for a given l. one of the main problem of the method is the stacking of satellites in host luminosity bins , to recover sufficient statistics . in each l bin , there is a significant scatter in halo mass , which prevents to derive a unique halo mass - luminosity relation ( more et al 2009a ) . to help raise the degeneracy , it is useful to compute the velocity dispersion by different weighting methods : the satellite weighted dispersion : @xmath7 , ( where the sum is over the @xmath8 central galaxies , identified by their @xmath9 number , who have each @xmath10 satellites ) , is biased towards higher values than the host weighted one : @xmath11 , and the scatter increases with luminosity ( more et al 2009b ) .
one other caveat is to discriminate against interlopers , the fraction of which depends on the host selection criteria ; the interlopers fraction has been estimated from 10% for extremely isolated hosts ( prada et al 2003 ) , to 30% or more , in mock catalogs ( van den bosch et al 2004 ) .
the difficulty is that interlopers are not uniformly distributed on the sky , but are also clustered ( e.g. chen et al 2006 ) . with the help of the mock simulations to interpret the data ,
a quiterobust result is that halo mass to luminosity
m / l decreases with l , and more specifically , the dispersion - luminosity relation depends on the radius it is estimated : @xmath12 at 120kpc and @xmath13 at 350kpc ( prada et al 2003 ) .
it is interesting to compare the results with those obtained through weak lensing ( hoekstra et al 2002 ) .
the slope of the dispersion - luminosity relation is very close to the tully - fisher slope 1/4 for spiral galaxies ( verheijen 2001 ) , and can be called a tully - fisher equivalent .
recently , klypin & prada ( 2009 ) have carried out a further satellite study from the sdss , selecting as hosts only red isolated galaxies , expected to be ellipticals or spheroids .
since the hosts are very isolated , there are only 1 or 0 satellite for each galaxy .
they compute the radial distribution of velocity dispersion , in three host luminosity bins , and find constraints on dark matter models ( @xmath0cdm or modified gravity mond ) , while fitting both the dispersion @xmath2 , and the radial density law of the satellites .
they claim that mond can not account for the observations .
the fits however have to include many free parameters , related to the velocity anisotropy of satellites , and therefore contain significant degeneracy .
the situation is quite similar to the velocity dispersion studies at smaller scales around elliptical galaxies .
the drop of @xmath2 derived from planetary nebulae was first interpreted as a possible dearth of dark matter ( romanowsky et al 2003 ) , while fits including a radially variable anisotropy of velocities @xmath14 reconciled the data with the cdm model ( dekel et al 2005 ) .
@xmath14 can vary between -@xmath15 ( purely tangential orbits ) , to 1 ( purely radial orbits ) , passing through 0 ( isotropy ) , and this is justified through galaxy mergers , since ellipticals are assumed to be the result of major mergers , or a succession of minor mergers .
we used the satellite kinematics data from klypin & prada ( 2009 ) to test the predictions of mond at large - scale around isolated galaxies .
it has been shown that when the anisotropy parameter radial variation is properly taken into account , mond gives a good fit of the data ( angus et al 2007 , tiret et al 2007 ) .
we also combined around a typical well - studied early - type galaxy ngc 3379 the various fits of velocity dispersion at three different scales : small - scale with stellar tracers , intermediate scale with planetary nebulae , and large - scale with satellites ( in the corresponding luminosity bin ) .
both cdm and mond models provide satisfying fits ( although cdm cores have to be assumed in the center , tiret et al 2007 ) .
the anisotropy is comparable to what is expected in cosmological simulations ( sales et al 2007 ) .
we present in figure 1 more fits , corresponding to different mass and luminosity intervals , not available before , together with the radial distribution of tracer density and velocity anisotropy .
for all these fits , a constant stellar mass to light ratio of @xmath16=4 has been adopted , corresponding to realistic populations for the red galaxies involved .
note that given the tully - fisher equivalent relation observed , i.e. @xmath17 , the fit can easily be generalised for a wide range of mass , since these outside regions are in the mond regime , where @xmath18 ( milgrom 1983 ) .
these fits are for isolated galaxies .
we might expect problems in groups and clusters , when large masses exist nearby .
they produce then an external field effect ( efe ) , that reduce the dark matter equivalent of mond ( although the efe has been successfully fitted in the milky way , wu et al 2007 ) . at even larger scales ,
the dark halo mass - to - luminosity depends on environment , and for groups on the crossing time .
small and large haloes have the largest m / l , while intermediate haloes , small groups with late - type galaxies have the lowest m / l of @xmath1990 ( tully 2005 ) .
the method of satellite kinematics is giving now more robust results , with the increased statistics of big surveys ( sdss , 2df ) .
however , still very different results can be found in the literature , according to the selection of primaries ( isolation criterium ) , and the elimination of interlopers .
the mass and radial dependence of @xmath20 is now derived , but with large uncertainties , due essentially to stacking problems , the mass being widely scattered in a given luminosity bin .
the modelisation involves numerous degrees of freedom , in the radial distribution of the velocity anistropy , essentially .
the shape of @xmath20 versus distance can be fit with appropriate @xmath14 both in @xmath0cdm and mond .
the generalisation to a large range of masses , is automatic if the tully - fisher equivalent relation is satisfied @xmath17 .
angus , g. w. , famaey , b. , tiret , o. , combes , f. , zhao , h. s. : 2008 mnras 383 , l1 brainerd t. g. , specian m. a. 2003 apj 593 , l7 chen , j. , kravtsov , a. v. , prada , f. et al . : 2006 , apj 647 , 86 dekel a. , stoehr , f. , mamon , g. a. et al . :
2005 , nature 437 , 707 erickson , l. k. , gottesman , s. t. , hunter , j. h. : 1987 , nature 325 , 779 hoekstra , h. , yee , h. k. c. , gladders , m. d. : 2002 , newar 46 , 767 klypin a. , prada f. : 2009 apj 690 , 1488 mckay , t. a. , sheldon , e. s. , johnston , d. et al : 2002 apj 571 , l85 milgrom m. : 1983 , apj 270 , 365 more , s. , van den bosch , f. c. , cacciato , m. et al .
: 2009a , mnras 392 , 801 more , s. , van den bosch , f. c. , cacciato , m. : 2009b , mnras 392 , 917 norberg , p. , frenk , c. s. , cole , s. : 2008 mnras 383 , 646 prada , f. , vitvitska , m. , klypin , a. : 2003 apj 598 , 260 romanowsky , a.j . , douglas , n.g . , arnaboldi , m. et al : 2003 science , 301 , 1696 sales l.v . ,
navarro , j. f. , lambas , d. g. et al . : 2007 , mnras 382 , 1901 tiret o. , combes f. , angus , g. w. , famaey , b. , zhao , h. s. : 2007 a&a 476 , l1 tully r. b. : 2005 apj 618 , 214 van den bosch f. c. , norberg , p. , mo , h. j. , yang , x. : 2004 mnras 352 , 1302 verheijen m.a.w . : 2002 , apj 563 , 694 wu , x. , zhao , h - s . , famaey , b. et al : 2007 , apj 665 , l101 zaritsky , d. , smith , r. , frenk , c. , white , s. d. m : 1993 , apj 405 , 464 zaritsky d. , white s.d.m . : 1994 , apj 435 , 599 zaritsky , d. , smith , r. , frenk , c. , white , s. d. m : 1997 , apj 478 , 39 | recently , the distribution of velocity dispersion as far as 400kpc around red isolated galaxies was derived from statistical studies of satellites in the sdss ( klypin & prada 2009 ) .
this could help to constrain dark matter models at intermediate scales .
we compare the predictions of different dm distributions , @xmath0cdm with nfw or cored profiles , and also modified gravity models , with observations . it is shown how the freedom in the various parameters ( radial distribution of satellites , velocity anisotropy , external field effect ) , prevents to disentangle the models , which all can give pretty good fits to the data . in all cases ,
realistic radial variations of velocity anisotropy are used for the satellites , and a constant stellar - mass to light ratio for the host galaxies .
address = observatoire de paris , lerma , 61 av .
de lobservatoire , f-75014 paris , france address = sissa , via beirut 4 , i-34014 trieste , italy |
it was realized already by a. bohr @xcite , who discussed the general properties expected for a full ( nn , pp and np ) components that t=1 pairing phonon is appropriate for the region around @xmath2ca . already then , the evidence including both energetics and transfer data .
@xcite was compelling regarding the major role played by the t=1 pairing in n = z nuclei .
these ideas were further developed by bs and broglia and culminated in the review article @xcite , where a very detailed analysis of isovector pairing vibrations was presented .
studies involving isospin effects have undergone a resurgence in recent years as such nuclei become more readily accessible .
moreover , near closed shells , the strength of the pairing force relative to the single - particle level - spacing is expected to be less than the critical value needed to obtain a superconducting solution , and the pairing field then gives rise to a collective phonon .
however , despite many experimental efforts , these predictions have not been confirmed yet .
macchiavelly et al .
, @xcite presented an experimental analysis of the pairing vibrations around @xmath16ni with emphasis on odd - odd nuclei .
their results clearly indicate a collective behavior of the isovector pairing vibrations .
the @xmath3 of such vibrations is estimated to be 0.8 mev @xcite for @xmath2ca . in a recently published article cedervall et al .
, @xcite obtained evidence for a spin - aligned neutron - proton paired phase from the level structure of @xmath17pd .
gezerlis et al .
, @xcite discussed also the possibility of mixed - spin pairing correlations in heavy nuclei .
the aim of the present work was to study the low - lying dipole strength distribution in @xmath2ca by using the ( @xmath1he , t ) reaction , which was extensively used earlier for the excitation of spin - isospin vibrational states in other isotopes .
some of our preliminary results was published recently @xcite .
the @xmath2ca(@xmath1he , t)@xmath2sc reaction has been studied earlier by schulz et al .
, @xcite at 28 mev bombarding energy and by loiseaux et al .
, @xcite at 30.2 mev bombarding energy using solid state telescopes and a magnetic spectrograph .
the energy resolution was 70 kev and 15 - 20 kev , respectively .
some of the @xmath18 , @xmath19 and @xmath20 proton - neutron multiplet states are identified and the effect of configuration mixing is discussed .
a similar experiment has been performed more recently by hansper et al .
, @xcite at 26.1 mev , using a magnetic spectrometer with an energy resolution of 15 kev .
correspondence of the observed @xmath2sc levels with the known t=1 states in @xmath2k and @xmath2ca are based on predictions provided by the isobaric multiplet mass equation .
the @xmath21ca(@xmath1he , t)@xmath21sc reaction was studied earlier by grewe et al.,@xcite at 420 mev bombarding energy with an energy resolution of about 40 kev .
up to about 9 mev some excited states relevant for the double @xmath22-decay were identified .
those levels were observed and investigated also in our present work and they are in good agreement with the levels , observed by grewe et al .
the spin - isospin excitation has been investigated earlier by tabor et al .
, @xcite in the @xmath2ca(@xmath1he , t)@xmath2sc reaction at 130 and 170 mev .
the angular distribution was measured for the suspected giant dipole resonance ( gdr ) structure .
the data are reasonably well described by a collective model calculation based on the goldhaber - teller model of the gdr .
some weaker l=1 resonances at 2 , 4 , 6 and 8 mev has also been observed .
however , their energy resolution of about 400 kev did not allow to study their structures .
the experiment was performed at the research center for nuclear physics , osaka university .
the energy of the @xmath1he beam of @xmath23 mev was achromatically transported to self supporting metallic @xmath2ca , @xmath24ca , @xmath25ca , @xmath21ca targets with thicknesses of @xmath26 mg / cm@xmath14 .
the typical beam current was @xmath28 na .
the energy of the tritons was measured with a magnetic spectrometer `` grand riden '' , using complete dispersion matching techniques @xcite .
the energy resolution was about @xmath29 kev .
the spectrometer was set at 0@xmath30 and 2.5@xmath30 with respect to the beam direction with an opening angle of @xmath3120 mrad horizontally and @xmath3120 mrad vertically defined by a slit at the entrance of the spectrometer .
a few typical triton spectrum is shown in the left side of fig . 1 . for the different ca targets indicated in the figures , showing preferentially the @xmath32l=1 , dipole transitions .
right side : examples for the angular distributions measured in the @xmath2ca(@xmath1he , t ) @xmath2sc reaction for @xmath32l=0 ( a ) , for @xmath32l=1 ( b ) and for some not identified ( c ) transitions , width=529 ] .
the spectra were analyzed using the program package : gaspan . in a given energy range
all peaks were fitted at the same time .
gaussian line shape with exponential tails and second order polynomials were used for describing the background .
the quality of the fit was always good .
the excitation energies of the ias s and a few well known excited states are used for determining the precise energy calibration .
we determined the precise level energies and intensities for each isotope .
the spectra were studied in eight distinct angular regions for all scandium isotopes .
the angular distributions were determined for each known , and new peaks .
a few examples for the angular distributions are displayed at the right hand side of fig
the lowest lying states in @xmath2sc are identified as members of the @xmath33 @xmath34 ( j@xmath35(j@xmath35 = @xmath36 ) multiplet . as the ( @xmath1he , t )
reaction at this bombarding energy excites preferentially the spin - flip states , in our case the 2@xmath37 state is excited the strongest , in which the spin of the proton and the neutron hole is parallel .
the isospin of such state is t=1 and t@xmath38=-1 .
such proton neutron multiplet has been observed also in the mirror nucleus of @xmath2sc namely in @xmath2k in which it is also the ground state multiplet with t=1 and t@xmath38=1 .
the t@xmath38=0 members of the isospin multiplet has been observed in @xmath2ca at 7.658 mev above the ground state .
such a shift can be explained by the coulomb energy difference .
the other states of the proton - neutron multiplet with j@xmath35= 3@xmath37 , 4@xmath37 and 5@xmath37 were also identified in @xmath2sc(t=-1 ) , in @xmath2ca(t=0 ) and in @xmath2k(t=1 ) @xcite .
all three multiplets turned out to be very similar .
the sequence of the j@xmath35 s are 4@xmath37 , 3@xmath37 , 2@xmath37 and 5@xmath37 .
the energy differences between the members of the bands agree(s ) within 1 - 2 kev . in @xmath2sc
no other multiplet states has been identified yet . however , using the strong similarity of the low - lying excited states in @xmath2k and @xmath2sc we may identify some additional multiplet states .
the next multiplet in @xmath2k is @xmath39 with l=2 and j@xmath35 s are 0@xmath40 ( 1644 kev ) , 2@xmath40 ( 1959 kev ) , 3@xmath40 ( 2260 kev ) and 1@xmath40 ( 2290 kev ) . at about the same excitation energy the @xmath41 ( l=1 , j@xmath35 = @xmath42 ) multiplet
was also identified .
as the @xmath43 is somewhat lower in @xmath44sc than in @xmath45k such a multiplet should be also lower in @xmath2sc than in @xmath2k .
the triplet state observed around 1.7 mev is a good candidate for the @xmath46 multiplet in @xmath2sc . this assignment is supported also by the angular distribution of the states .
the dipole strengths distributions deduced from our experimental data for @xmath2sc , @xmath24sc , @xmath25sc , and @xmath21sc are compared in fig .
2 . sc ( a ) , @xmath24sc ( b ) , @xmath25sc ( c ) , and @xmath21sc ( d ) as a function of the excitation energy.,width=453 ]
concerning the dipole strengths distributions shown in fig .
2 two observation can be made : 1 . there is a relatively strong peak at about 10 mev in each distributions , 2 .
some periodic structure of the distribution is showing up , especially for @xmath2sc .
+ both features are very common for all isotopes , which suggests the presence of some core excitations in @xmath2ca .
both features are very common for all isotopes , which suggests the presence of some core excitations in @xmath2ca . in order to understand the experimental results relativistic rpa ( rrpa ) calculations
have been performed with nl3 @xcite and dd - me2 @xcite interactions .
the results are shown in fig .
3 . rrpa predicts low - lying strength with nearly periodic peaked structure caused by the dipole isospin - flip and spin - isospin - flip transitions governed by the pion and effective rho - meson exchange interactions .
the obtained strength , however , shows no enhancement of the strength at lowest energies .
thus , it is expected that correlations beyond rrpa can be responsible for the observed enhancement .
it was realized that the periodic structure observed in the dipole strengths distribution can be associated with the multiparticle - multihole @xmath47 states observed previously in @xmath2ca , and also shown in the left part of fig .
ca(@xmath1he , t)sc reaction at @xmath48 . a folded spectrum ( with fwhm=500 kev )
is shown in red .
the periodic structure of the distribution is clearly visible .
the position of the strong 2@xmath37 transition of the ground state multiplet as well as the results of their coupling to the lowest lying 0@xmath40-states in @xmath2ca are marked .
right part : part of the level scheme of @xmath2ca showing only the 0@xmath40 levels , the level density of the 0@xmath40-states and the shell model prediction for the multiparticle - multihole configurations @xcite.,width=453 ] coupling the 2@xmath37 state of the ground state multiplet ( which is the strongest channel in the ( @xmath1he , t ) rection ) to the different low - lying 0@xmath40 states in @xmath2ca , the centroids of the bumps at 4,6 and 8 mev can nicely be reproduced .
the density of the states increases rapidly above 8 mev .
the distribution of the 0@xmath40 states is also shown in the right part of the figure .
it has a definite peak at around 10 mev .
coupling this peak to the 2@xmath37 state , one gets the 10 mev peak in the sdr distribution .
such low - lying 0@xmath40 states in @xmath2ca can also be considered as parts of the giant monopole resonance ( gmr ) .
the gmr was investigated by yangblood et al . , @xcite and indeed observed a bump around 10 mev , although the centroid of the gmr was found to be at 19 mev .
we expect also coupling of the 2@xmath37 state to the gmr .
the high energy dipole stench distribution was investigated by gaarde et al .
@xcite in the @xmath2ca(p , n ) reaction at 200 mev bombarding energy and observed such a strong dipole peak at about 22 mev . according to the shell model calculations of m. sakakura et al.,@xcite the energy of the 4,6 and 8 particle - hole states is also shown in fig .
4 . such multiparticle - multihole configurations might be associated to monopole multiphonon states as well .
we have a proton - neutron pair connected to such multiphonon states .
the resulting 2@xmath37 states are also coupled to hundreds of other 2@xmath37 states ( we are dealing with an odd - odd nucleus ) , which result in the observed spreading of their strengths .
similar periodic structure is expected in the dipole strengths distribution of @xmath2k excited in the @xmath2ca(n , p ) reaction . for the n = z @xmath2ca nucleus the sum rule for the @xmath49 and @xmath50 strengths reduces to @xmath51 @xcite .
so the cross section of the @xmath2ca(p , n ) and the @xmath2ca(n , p ) reactions should be the same .
for isospin symmetry considerations one expects the same strengths distribution as well for the ( p , n ) and ( n , p ) type reactions . unfortunately , most of the ( n , p ) reactions were performed at low bombarding energy where the spin - isospin exciations were suppressed . the only reaction , which was studied at intermediate energy ( 152 mev ) was performed by maesday et al . ,
@xcite , but the energy resolution was about 3 mev , which smeared out the structure of the spin dipole strengths distribution .
they observed only one broad peak , which was identified as the gdr . *
acknowledgments * the authors acknowledge the rcnp cyclotron staff for their support during the course of the present experiment . this work has been supported by the hungarian otka foundation no .
k106035 and by the loewe program of the state of hesse ( helmholtz international center for fair ) .
99 a. bohr , b.r .
mottelsson , nuclear structure , vol .
2 , benjamin - new york , p. 645
o. nathan , proc . of dubna symposium on nuclear structure , iaea ,
p.191 ( 1969 ) .
bes et al . , phys . rep .
34c ( 1977 ) 1 , and references therein .
macchiavelly et al . , phys .
b 480 , 1 ( 2000 ) .
b. cederwall et al . ,
nature 469 , 68 ( 2011 ) .
a. gezelris , g.f .
bertsch , y.l .
luo , phys .
106 , 252502 ( 2011 ) .
l. stuhl et al . , acta physica polonica b 42 , 667 ( 2011 ) .
n. schulz w.p .
alford , a. jamshidi nucl .
a162 , 349 ( 1971 ) .
loiseaux et al . , phys . rev .
c4 , 1219 ( 1971 ) .
v.y . hansper et al . , phys .
c61 , 028801 ( 2000 ) .
grewe et al . , phys . rev .
c76 , 054307 ( 2007 ) .
tabor et al . , phys .
c25 , 1253 ( 1982 ) .
h. fujita et al . , nucl .
instr . and meth .
a484 , 17 ( 2002 ) .
cameron and b. singh , nucl .
data sheets , 102 293 ( 2004 ) .
lalazissis , j. knig , and p. ring , phys . rev .
c 55 , 540 ( 1997 ) .
lalazissis , t. niksic , d. vretenar , and p. ring , phys .
c 71 , 024312 ( 2005 ) .
d. h. youngblood , y .- w .
lui , and h. l. clark , phys .
c55 , 2811 ( 1987 ) . c. gaarde et al .
a369 , 258 ( 1981 ) .
m. sakakura , a. arima and t. sebe , phys .
b61 , 335 ( 1976 ) .
a. krasznahorkay et al . , phys .
82 , 3216 ( 1999 ) .
measday , j.n .
161 , 1071 ( 1967 ) . | high resolution experimental data has been obtained for the + @xmath0ca(@xmath1he , t)sc charge exchange reaction at 420 mev beam energy , which favors the spin - isospin excitations .
the measured angular distributions were analyzed for each state separately , and the relative spin dipole strength has been extracted for the first time .
the low - lying spin - dipole strength distribution in @xmath2sc shows some interesting periodic gross feature .
it resembles to a soft , dumped multi - phonon vibrational band with @xmath3= 1.8 mev , which might be associated to pairing vibrations around @xmath2ca . * soft spin dipole giant resonances in @xmath2ca * l. stuhl@xmath4 , a. krasznahorkay@xmath4 , m. csatls@xmath4 , t. marketin@xmath5 , e. litvinova@xmath6 , t. adachi@xmath7 , a. algora@xmath8 , j. daeven@xmath9 , e. estevez@xmath8 , h. fujita@xmath7 , y. fujita@xmath10 , c. guess@xmath9 , j. gulys@xmath4 , k. hatanaka@xmath7 , k. hirota@xmath7 , h.j .
ong@xmath7 , d. ishikawa@xmath7 , h. matsubara@xmath7 , r. meharchand@xmath9 , f. molina@xmath8 , h. okamura@xmath7 , g. perdikakis@xmath9 , b. rubio@xmath8 , c. scholl@xmath11 , t. suzuki@xmath7 , g. susoy@xmath12 , a. tamii@xmath7 , j. thies@xmath13 , r. zegers@xmath9 , j. zenihiro@xmath7 + 0.5 cm @xmath4 inst . of nucl .
res .
( atomki ) , p.o .
box 51 , h-4001 debrecen , hungary , + @xmath14 gsi 64291 darmstadt , germany + @xmath1 phys .
dept .
univ . of zabreb , 10000 zagreb , croatia + @xmath15 inst .
fr theor .
phys .
goethe univ .
, 60438 frankfurt am main , germany + @xmath7 rcnp , osaka university , ibaraki , osaka 567 - 0047 , japan + @xmath8 ific , csic - universidad de valencia , 46071 valencia , spain + @xmath9 nscl , michigen state university , east lansing , michigan 48824 , usa + @xmath10 department of physics , osaka university , toyonaka , osaka 560 - 0043 , japan + @xmath11 institut fr kernphysik , universitt zu kln , 50937 kln , germany + @xmath12 istanbul univ .
, faculty of sci .
, phys .
dept .
, 34134 vezneciler , istanbul , turkey + @xmath13 institut fr kernphysik , universitt mnster , 48149 mnster , germany |
structural components of galaxies like bars , bulges , ovals and bar - related resonance rings are among the key factors for evaluating when present - day galaxies formed and how they evolve over time .
bulges in early - type disk galaxies are generally assumed to be old , formed through hierarchical clustering , but they might also be more recent structures assembled by internal dynamical processes in galaxies . in favorable conditions these processes might lead to transformation of spirals into s0s , through gas stripping in the halo and the disk ( bekki , warrick @xmath5 yashuro 2002 ) , or due to strong and sustained star formation in the central regions of galaxies ( kormendy @xmath5 kennicutt , hereafter kk2004 ) . 0.25 cm in these secular evolutionary processes the relative mass of the bulge increases .
in case of gas stripping the bulges should have older stellar populations than the disks and their surface brightness profiles are expected to resemble more the @xmath6 law than the exponential function .
however , due to a small amount of gas in s0s , star formation is not expected to be the dominant mechanism for the growth of the bulge , unless exceptionally strong accretion of external gas is taking place in these galaxies .
bulge - like mass concentrations can form also through resonant scattering or through bending instabilities in the central regions of primary bars ( pfenniger 1984 ; raha et al .
1991 ) , in which case these so called `` pseudo - bulges '' should have similar stellar populations as the underlying bars . in order to evaluate the possible role of these processes in galaxy evolution , the properties of bars , ovals and bulges need to be systematically studied in the spiral / s0 divide , e.g. through sa , s0/a , to s0@xmath7 , s0@xmath8 types .
0.25 cm secondary processes in galaxy evolution are generally tied to the growth of a bulge in one way or another , but an alternative approach was also discussed by van den bergh ( 1979 , 1998 ) , who outlined the ddo system of classification in terms of a nurture theory . in this classification system , s0s form a sequence parallel to spirals with types s0a , s0b , and s0c being the s0 analogues of sa , sb , and sc galaxies . a difficulty with this approach has been that no s0c galaxies have been found so far , that is , clear s0 galaxies showing a bulge to total luminosity ratio similar to an sc galaxy . however , this might be a bias due to the fact that application of too simple analysis methods can easily hide the real nature of bulges , particularly in s0s which have rich morphological structures .
0.25 cm to our knowledge no previous systematic morphological study of galaxies has been carried out in the spiral / s0 divide in the near - infrared . also ,
although 2-dimensional decompositions have become a standard , even in recent studies ( de souza , gadotti @xmath5 dos anjos 2004 , hereafter sga2004 ; christlein @xmath5 zabludoff 2004 ) multiple structures have not been taken into account .
the current study is a third paper in a series where morphological properties of a sample of 170 early - type disk galaxies are studied .
the description of the sample and the first results are presented in paper i by laurikainen , salo @xmath5 buta ( hereafter lsb2005 ) , and in paper ii by buta et al .
( 2006 , in press ) .
in paper i we found , for a subsample of 24 s0-s0/a galaxies , that the @xmath4 flux ratios are significantly smaller , and the bulges more disk - like , than generally assumed for their morphological type . in paper
ii we investigate in detail the fourier properties of early - type bars and ovals , and identify bars , which are well described in terms of single or double gaussian components of the relative fourier amplitudes of density . in this study
we analyze the morphological properties of 15 southern galaxies of the de vaucouleurs revised hubble types s0@xmath8 to sa , mainly s0s , with a large range of apparent bar strengths .
for these galaxies we use @xmath0-band images to derive the orientation parameters from the elliptical isophotes , apply fourier methods for calculating non - axisymmetric forces , and perform 2d multicomponent structural decompositions .
the multicomponent decompositions are compared with more simple bulge / disk decompositions and with 1d decompositions for the same galaxies
. 0.25 cm
0.25 cm this is a sub - sample of 15 galaxies of the near - infrared s0-survey ( nirs0s ) , which was originally drawn from the third reference catalog of bright galaxies ( de vaucouleurs et al .
1991 , rc3 ) having the following selection criteria : -3 @xmath9 @xmath10 @xmath11 2 , @xmath12 @xmath11 12.5 mag and logr@xmath13 @xmath11 0.35 , yielding a sample of 170 galaxies .
the final sample consists of 190 galaxies , because 20 late - type ellipticals classified as s0s by sandage , were also added .
this is for taking into account also the misclassified es in rc3 .
the original sample was selected to be similar in size with the ohio state university bright galaxy survey ( osubgs ; ( eskridge et al .
2002 ) for spirals .
the properties of the sample are described in more detail in paper ii .
0.25 cm the observations were carried out in december 2004 using the sofi instrument , attached to the 3.5 m new technology telescope ( ntt ) at eso .
high resolution @xmath0-band images ( 0.29 arcsec / pix ) were obtained for all galaxies , and for five of them also @xmath14-band images .
the total on - source integration time was 1800 - 2400 sec in the @xmath0-band , and 1200 sec in the @xmath14-band , mostly taken in snap - shots of 3 - 6 seconds .
owing to the high sky brightness in the near - infrared , the sky fields were taken alternating between the galaxy and the sky field in time intervals of 1 min : in the object a random dithering within a box of 20 arcsec was used , whereas the sky fields were randomly selected carefully avoiding bright foreground stars .
the observations were obtained in good atmospheric conditions , the nights being mainly photometric .
the mean stellar full width at half maximum ( fwhm ) is 0.72 arcseconds in the @xmath0-band .
the sample and the seeing conditions are shown in table 1 , where stdev is the standard deviation of the mean in the sky background . the galaxy classifications in column 3 are from rc3 .
the adopted morphological type in column 2 is mainly from the de vaucouleurs atlas of galaxies by buta , corwin @xmath5 odewahn ( 2007 , hereafter bco2007 ) , or the classification is judged by buta using the images in sandage @xmath5 bedke catalog ( 1994 ) .
for ngc 3706 and eso 208-g21 , the classification is based on the @xmath0-band images from this study .
0.25 cm the images were processed using iraf package routines .
the reduction steps consisted of sky subtraction , flat - fielding of the individual difference images , removal of foreground stars , and transposing the images to have north up and west on the right .
dome flats were used , in which the shade patterns due to the scattered light were corrected .
the uneven illumination of the screen was also corrected in the flat - fields using a correction frame , created from multiple observations of one standard star .
the standard star observation was used to create a surface , which was then normalized to unity . for these corrections
iraf scripts offered by eso were used .
the images were cleaned of foreground stars by finding the stars with daofind , and then using a stellar psf to remove the stars .
any residuals left by the psf fitting were cleaned using iraf routine imedit . for five of the galaxies @xmath15 color maps
were also constructed .
the images and the color maps are shown in figure .
1 . all direct images in figure 1 are in units of mag arcsec@xmath16 and are generally displayed from @xmath17 = 12.0 to 22.0 mag arcsec@xmath16 .
color index maps display colors in the range 0.5 @xmath11 @xmath15 @xmath11 1.5 .
0.25 cm the elliptical isophotes were derived mainly in the @xmath0-band using the ellipse routine in iraf .
we calculate the radial profiles of the position angles @xmath18 and the ellipticity @xmath19 = 1 - @xmath20 , where @xmath20 is the minor - to - major axis ratio ( see fig .
the profiles were calculated both in linear and in logarithmic scale : the linear scale for deriving the orientation parameters of the outer disks , and the logarithmic scale for identifying the structural components particularly in the inner regions of the galaxies .
the isophotal fits were made to the original images .
when available , @xmath14-band images were used to estimate the orientation parameters , because they were generally deeper than the @xmath0-band images . in the isophotal fitting deviant pixels above @xmath21 were rejected .
the parameters @xmath18 and @xmath20 are listed in table 2 , calculated as mean values in the radial range indicated in the table .
the uncertainties are standard deviations of the mean in the same radial range . also shown
are the orientation parameters given in rc3 , and for some of the galaxies values from other sources as well .
the values used in this study are shown in boldface . in the analysis orientation parameters derived in this study
are used , except for the galaxies ngc 3081 , ngc 1512 and ngc 1326 . for the first two galaxies we use kinematical estimates of @xmath18 obtained by buta @xmath5 purcell ( 1998 ) and buta ( 1988 ) , respectively , and for ngc 1326 the photometrical estimates by buta et al .
( 1988 ) is used . 0.25 cm a comparison to rc3 shows large deviations in the orientation parameters particularly for the galaxies ngc 1079 ( @xmath18 ) , ngc 1317 ( @xmath20 and @xmath18 ) , ngc 1512 ( @xmath18 ) , ngc 1533 ( @xmath20 and @xmath18 ) and ngc 3081 ( @xmath20 , @xmath18 ) . however , our measurements are generally in good agreement with those measurements given in the literature , where deep optical or ir images are used .
this is the case for ngc 1512 , for which kuchinsky et al .
( 1998 ) obtained @xmath18=56@xmath22 . for ngc 3081 buta @xmath5 purcell ( 1998 )
give photometrically estimated values of @xmath18=86@xmath22 and @xmath20=0.387 .
also , for ngc 1351 our measurements for the orientation parameters are in agreement with those obtained by garcia - gomez et al .
0.25 cm for estimating tangential forces induced by non - axisymmetric structures in galaxies we use a gravitational bar torque approach ( combes @xmath5 sanders 1981 ; quillen , frogel @xmath5 gonzalez 1994 ) , first applied to galaxy samples by ( buta @xmath5 block 2001 ; laurikainen @xmath5 2002 ) . the polar method we use has been described by laurikainen @xmath5 salo ( 2002 ) , salo et al .
( 2004 ) , and laurikainen , salo @xmath5 buta ( 2004 , hereafter lsb2004 ) ( see also salo et al .
the gravitational potential is inferred from the two - dimensional @xmath0-band image by assuming that the near - ir light distribution traces the mass and that the mass - to - luminosity flux ratio is constant .
it is assumed that the vertical profile is exponential and that the vertical scale height depends on the morphological type .
deprojected images were used , which were constructed by first subtracting off the bulge model , deprojecting only the disk light , and then adding back the bulge as an assumed spherical component ( see section 5 also for more sophisticated decompositions ) .
the derived gravitational potentials are used to calculate maps of the maximum ratio of the tangential force ( @xmath23 ) to the azimuthally averaged radial force ( @xmath24 ) , as well as a radial profile of the maximum relative tangential force at each distance ( see fig .
3 ) , @xmath25 where @xmath26 denotes the azimuthally averaged axisymmetric force for each @xmath27 . the maximum value of this profile , @xmath28 at a radius @xmath29 , gives a single measure of _ bar strength _ for these early - type galaxies where spiral arm torques are generally negligible . in the calculation of the gravitational potential we use
azimuthal fourier decompositions in radial annulae , typically using even components with azimuthal wave number @xmath30 20 .
the gravitational potential is then calculated using fft in the integration over azimuth , and direct summation in radial coordinates .
an important advantage of this method , compared to the more straightforward cartesian integration ( `` quillen s method '' ) is the suppression of spurious force maxima in the outer disk arising due to noise .
the other advantage is that it gives simultaneously also the relative mass of the bar , as estimated from the maxima of @xmath31 = 2 and @xmath31 = 4 amplitudes of density ( @xmath32 and @xmath33 , respectively ) .
the measurements are shown in table 3 , where the bar lengths are also given .
bar length is identified within the region where the phase of the @xmath31 = 2 amplitude is maintained nearly constant .
if no bar length is given , the parameters indicate the properties induced by other non - axisymmetric components like ovals or inner disks .
the errors are formal errors related to @xmath28 measurements in the four image quadrants .
it is worth noticing that a peak in a @xmath34-profile should not be automatically interpreted to indicate a presence of a bar .
for example , in ngc 3100 strong tangential forces are detected at @xmath27 @xmath11 40 , although there is no bar in this galaxy . as discussed in section 5 ,
this galaxy is dominated by a large oval , which appears as a non - axisymmetric structure in the fourier analysis .
because of the high uncertainty in the position angle of ngc 1512 , no @xmath28 value for this galaxy is given in table 3 : the kinematic estimate by buta ( 1988 ) gives @xmath18 = 83@xmath22 while we measure @xmath18 = 46@xmath22 . 0.25 cm
0.25 cm 0.25 cm we use a 2-dimensional ( 2d ) multicomponent decomposition code , originally presented by lsb2005 ( see also lsb2004 ) , where the code was also tested . in this study
we use also a one - dimensional ( 1d ) version of the code ( laurikainen @xmath5 salo 2000 ) , which can be efficiently used in conjunction with the 2d code . in the 2d multicomponent code
all components use generalized elliptical isophotes ( atthanassoula et al .
1990 ) , the isophotal radius being described by the function : @xmath35 the isophote is boxy when the shape parameter @xmath36 0 , disky when @xmath37 0 , and purely elliptical when @xmath38 = 0 .
circular isophotes correspond to @xmath38 = 0 and minor - to - major axial ratio @xmath20 = 1 .
the disks are described by an exponential function : @xmath39,\ ] ] where @xmath40 is the central surface density of the disk , and @xmath41 is the radial scale length of the disk .
the radius @xmath42 is calculated along the disk plane defined by the assumed @xmath18 and @xmath20 ( for disk c = 0 ) .
the surface brightness profiles of the bulges are described by a generalized sersic s function : @xmath43,\ ] ] where @xmath44 is the central surface density , @xmath45 is the scale parameter of the bulge , and @xmath46 = @xmath47 determines the slope of the projected surface brightness distribution of the bulge ( @xmath1 = 1 for exponential , and @xmath1 = 4 for the @xmath6-law ) .
the parameter @xmath48 is the isophotal radius defined via the parameters @xmath49 , @xmath50 and @xmath51 , where the subscript stands for the bulge . in the case of an elliptical bulge
, @xmath51 is its major axis position angle measured counter clockwise from north in the sky plane .
sersic s function can be applied also for the other non - axisymmetric components , or alternatively the ferrers function is used , described as : 0.25 cm @xmath52 where @xmath53 is the central surface brightness of the bar , @xmath54 is the bar major axis length , and @xmath55 is the exponent of the bar model defining the shape of the bar radial profile . the isophotal radius ( @xmath56 is defined via parameters @xmath57 , @xmath58 , @xmath59 . as in the 2d version ,
also in the 1d version the bulges are described by a sersic s function .
0.25 cm 0.25 cm the fitting was made in flux units in the plane of the sky .
depending on the choice of the weighting function more weight can be assigned to the pixels in the central parts of the galaxies where the flux level is high , or to the outer regions , where the number of pixels is high . in this study
we generally use @xmath60 , where @xmath61 is the model flux in pixel @xmath62 .
however , it was shown by lsb2005 that our method is not sensitive to the choice of the weighting function . in order to account for the effects of seeing
, the model is convolved with a gaussian psf using a fwhm measured for each image .
0.25 cm due to the large number of free parameters in the 2d approach the fitting is time consuming . in order to speed the process a good fit
was first searched for the images rebinned by a factor of 4 , which results were then used as initial parameters for more sophisticated decompositions performed for the original images . generally the decompositions were started by finding @xmath41 . in cases where the presence of an extended disk was not immediately obvious , spiral arms or rings were used to confirm the presence of the underlying disk . in some cases they were visible only after subtracting the bulge or disk model .
a few images were too noisy for finding @xmath41 in a reliable manner with the 2d code , in which case the 1d code was used by fitting an exponential to the outer parts of the disk .
0.25 cm three different kinds of 2d decompositions were made : ( 1 ) _ multicomponent _ ( @xmath63 = 7 ) , ( 2 ) _ 3-component _
( typically bulge / disk / bar ) and ( 3 ) _ 2-component _ ( bulge / disk ) decompositions . additionally , _
simple 3-component _ decompositions were made by assuming that the bulges have spherical light distributions .
such decompositions are required for deprojecting the images , because it is difficult to make a 3-dimensional bulge model allowing also flattening or triaxiality for the bulge .
as not all bulges are spherical , some caution is needed while interpreting the innermost structures of the deprojected images .
the best - fitting multicomponent decompositions are compiled in table 4 and illustrated in figure 4 .
the numbers 1 , 2 and 3 in the title line in the table generally indicate the primary , secondary and tertiary components .
the component 1 is in most cases the primary bar , but in non - barred galaxies it can be an oval or an inner disk .
the component 2 generally refers to the secondary bar or oval , and the component 3 to a large tertiary bar or oval .
the bulge - to - total ( @xmath4 ) flux ratio can be calculated either using the disk - component extrapolated to infinity , or limiting to the fitted region alone : in this study the non - extrapolated disk is used .
0.25 cm to be sure that we were fitting real structures , a preliminary identification of bars , ovals and disks was made by inspecting ( 1 ) the original images and the @xmath64 color maps , 2 ) the radial profiles of the orientation parameters , and 3 ) the profiles of the fourier amplitudes of density . in some cases also the bulge- or disk model - subtracted images were inspected after - wards .
the non - axisymmetric structures appear as bumps in the @xmath19-profiles , provided that their surface brightnesses are high enough compared to that of the underlying disk .
also , the position angles show isophotal twists in the bar regions .
the even fourier amplitudes of density are particularly sensitive for detecting the non - axisymmetric structures in galaxies .
the identifications of the morphological components and the ring dimensions are shown in tables 5 and 6 . as in table 4 , also in table 5 the subscripts 1 and 2 refer to the primary and secondary bars / ovals . following kk2004
we adopt the terminology that the flat inner structure is called a `` lens '' if the morphological type is s0-s0/a , and an `` oval '' when the galaxy has a later morphological type
. a lens has been originally defined by kormendy ( 1979 ) as a feature showing a shallow brightness gradient interior to a sharp edge , whereas an oval was defined as having a broad elongation in the light distribution .
it seems that in some cases the division according to the morphological type is artificial .
s0s can actually have both ovals and lenses .
however , as the discrimination between ovals and lenses is not straightforward , we keep the present terminology . 0.25 cm 0.25 cm in the following we discuss the morphological properties of the individual galaxies , which in early - type disk galaxies are generally rich . detecting these structures
depends not only on the wavelength , but also on the image resolution and the integration time used .
we use deep , high resolution near - ir images , but in spite of that it is sometimes difficult to identify the structural components , particularly the bulges . in order to illustrate this also some details of the decompositions
are given .
in most cases it was not possible to use any semi - automatic procedure to find reliable solutions .
0.25 cm * eso 208-g21 * : eso 208-g21 is a bulge - dominated system with an elongated inner structure .
most probably for this reason , ( corwin , de vaucouleurs @xmath5 de vaucouleurs 1985 ) classified it as e4 , whereas in rc3 the inner structure was interpreted to be a weak bar / oval , leading to the classification of sab0@xmath8 .
martel ( 2004 ) found that the inner structure is a dusty inner disk , which interpretation was based on the analysis of high resolution @xmath65-images .
the disk is clearly visible in our unsharp masked image ( fig .
5 ) and it can be identified also in the ellipticity profile as a peak , followed by a decline up to the radius where the dominance of the bulge makes the profile flat .
the ellipticity profile we find is very similar to that obtained previously by scorza et al .
( 1998 ) , who also made a structural decomposition for this galaxy . as they use a different method than we do
it is interesting to compare the results : in both cases a solution was found in which the galaxy has a small exponential inner disk , embedded in a large massive bulge .
the @xmath4 = 0.97 we found is very similar to @xmath4 = 0.92 obtained by scorza et al . in the @xmath66-band
( they give @xmath67 = 0.09 ) .
the decomposition by scorza et al . is based on an iterative process where the disk is first subtracted and the isophotes of the remaining bulge component are inspected .
the procedure is repeated until perfect elliptical isophotes remain .
it is worth noticing that for this galaxy our 1d and 2d codes give very similar solutions .
obviously this galaxy has no outer disk , and the bulge is strongly flattened .
we classify this galaxy as e(d)5 , thus confirming the earlier classification by kormendy @xmath5 bender ( 1996 ) .
0.25 cm * ngc 1079 * : the morphology of ngc 1079 has been previously studied by buta ( 1995),jungwiert , combes @xmath5 axon ( 1997 ) , sga2004 , who also made a 2d decomposition for it , and by knapen et al .
( 2006 ) , who found a strong nuclear peak in h@xmath68 surrounded by rather faint circumnuclear patchy emission .
buta ( 1995 ) recognized ngc 1079 as a galaxy showing one of the largest examples of a feature called r@xmath69r@xmath70 , a double outer ring / pseudoring morphology linked to an outer lindblad resonance ( buta @xmath5 crocker 1991 ) .
this feature includes a well - defined outer ring from which two conspicuous spiral arms emerge .
sandage @xmath5 brucato ( 1979 ) also recognized ngc 1079 as an early - type spiral .
the bar in ngc 1079 was recognized by jungwiert , combes @xmath5 axon as a large peak at @xmath27 = 30 arcsec in the ellipticity profile , while a smaller peak at @xmath27 = 17 arcsec was attributed to a triaxial bulge .
we see both peaks in our @xmath19-profile as well .
the axis ratio of the disk ( @xmath20 = 0.59 ) deviates from the axis ratio of the bulge ( @xmath20 = 0.68 ) , confirming that they are separate components
. 0.25 cm in our decomposition we first determined @xmath41 , which was then kept fixed
. finally the bulge , disk and the bar were fitted , but for the ansae - type morphology of the bar the solution depends slightly on the parameters of the bar : it is possible to use @xmath71 2 , or @xmath1 = 2 together with the shape parameter that deviates from purely elliptical isophotes .
depending on this choice we found @xmath4 = 0.20 - 0.25 .
our solution converges only if flattening of the bulge is taken into account .
the bar has a high surface brightness compared to that of the underlying disk .
apparently an inner ring surrounds the bar , seen as a residual after subtracting the model image from the observed image .
0.25 cm our solution for this galaxy is very different from the 2d bulge / disk solution obtained by sga2004 .
they found a large de vaucouleurs type bulge ( @xmath4 = 0.65 , @xmath1 = 3.82 ) dominating the whole surface brightness profile , and a very weak exponential inner disk , extending to the radius of the bar . in their fit the disk at the edge of
the bar is about 5 mag weaker than the bulge . sga2004
made the decomposition in the @xmath72-band , which extends approximately to the same radial distance as our @xmath0-band image .
for comparison we made also a 2d bulge / disk decomposition ( thus omitting the bar ) , which gives @xmath4 = 0.57 ( see fig .
6 , right upper panel ) .
this is near the @xmath4-ratio obtained by sga2004 , except that the bulge is more exponential ( @xmath1 = 2.7 ) .
however , we stress that in this solution the flux of the bar goes erroneously to the bulge .
a similar solution was found by us using the 1d - method .
this resembles the solution by sga2004 even more , although the disk we find is still several magnitudes brighter than that obtained by sga2004 .
it is also possible to find a solution in which the surface brightness profile is fitted by a single sersic s function with @xmath1 = 3.3 ( fig .
6 , middle upper panel ) , but in that case the fit is not good , and particularly , the model image is very different from the original image . 0.25 cm * ngc 1317 * : ngc 1317 has been previously reported to be a double barred galaxy , both in the optical ( schweizer 1980 ; wozniak et al .
1995 ) , and in the near - ir ( mulchaey , regan @xmath5 kundu(1997 ) .
the galaxy has also a nuclear ring , which is actually a double ring / pseudoring evident particularly in blue light ( schweizer 1980 ; papovich et el . 2003 ) and in h@xmath68 ( crocker , baugus @xmath5 buta 1996 ) .
the galaxy is nearly in face - on orientation . for the disk rc3
gives a larger ellipticity than found in this study , most probably because the outer isophotes are elongated towards the nearby companion .
0.25 cm the secondary bar is well visible in our @xmath0-band image , and the double nuclear ring is now detected for the first time also in the near - ir .
the outer structure is dominated by a large oval with no clear sign of a primary bar .
however , after subtracting the exponential disk a very weak classical bar becomes visible ( see fig .
previous evidence for the primary bar comes from the @xmath19-profiles , which is largely associated to the large oval .
the secondary bar lies within a bright nuclear lens ( first noticed by schweizer 1980 ) and is completely detached from the nuclear ring / pseudoring .
together , the secondary bar and its lens appear as a bump in the radial surface brightness profile . in the decomposition
only one primary component was fitted . for the complex structure of this galaxy
the decomposition was made in small steps .
the parameters for the disk and the main bar / oval were found first , after which they were fixed . then the secondary bar was found and fixed , and finally the bulge - component was found . in the central part of the galaxy the secondary bar is dominating over the central oval and
therefore it is also the component that the fitting procedure finds . in the final solution
all components ( except the shape parameter of the bar ) were left free for fitting .
the solution we find is reasonable , although it leaves some residuals , mainly because the bars and ovals can not be fitted separately .
0.25 cm the double nuclear ring in the @xmath15 color map shows a presence of dust , which is consistent with the observation that the ring is presently forming stars .
the ring and the secondary bar are also well visible in the @xmath0-band image after subtracting the bulge ( fig .
the ring has a patchy morphology , possibly indicating the presence of red super giants .
there are dust lanes also in the central regions of the galaxy , not aligned with the secondary bar .
0.25 cm * ngc 1326 * : this is a well studied galaxy previously reported to have a primary bar , a possible nuclear bar , as well as inner and outer rings . the outer ring is an exceptionally bright example of the outer lindblad resonance subclass r@xmath69 , characterized by an approximate alignment perpendicular to the bar and by slight dimpling in its shape near the bar axis ( buta @xmath5 crocker 1991 ) .
the nuclear ring is also an exceptional star - forming feature that has been well studied from both ground and space ( buta @xmath5 crocker 1991 ; garcia 1993 ; storchi - bergman et al .
1996 ; fernandes , storchi - bergmann @xmath5 schmitt 1998 ; buta et al .
2000 ) the inner bar and the inner and outer rings are visible in our @xmath0-band image , and the nuclear ring is evident also in the @xmath15 color map .
we detect a faint secondary bar inside the nuclear ring , which is best visible in the @xmath0-band image after subtracting the bulge model ( fig . 5 ) .
the possible double - barred nature of ngc 1326 was previously discussed by de vaucouleurs ( 1974 ) , buta @xmath5 crocker ( 1991 ) , wozniak et al .
( 1995 ) and erwin ( 2004 ) , but it has been difficult to verify owing to strong extinction inside the nuclear ring . the best evidence presented previously was by erwin ( 2004 ) , but the resolution in our @xmath0-band image is sufficient now to confirm it .
the @xmath15 color index map in figure 1 shows that dust still has some impact in the near - ir bands inside the nuclear ring .
the ring is red in this map which suggests the presence of red super - giants .
0.25 cm in our decomposition for ngc 1326 we fitted besides the bulge and the disk , also two bars .
this solution converges only if flattening of the bulge is taken into account .
models for the individual components were found in small steps and finally all components were fitted simultaneously . the bulge in this galaxy
was found to be quite prominent ( @xmath4 = 0.34 , @xmath1 = 3.0 ) but much less luminous than argued by sga2004 .
they found a very weak extended disk embedded inside a large de vaucouleurs type bulge dominating the whole surface brightness profile , leading to @xmath4 = 0.63 and @xmath1 = 4.9 .
for comparison we made also bulge / disk / bar and bulge / disk decompositions , the former leading to @xmath4 = 0.58 ( and @xmath1 = 3.1 ) and the latter to @xmath4 = 0.42 ( see also fig .
however , in the bulge / disk decomposition the flux of the bar / oval goes erroneously to the bulge .
it is also worth noticing that the b / t in the bulge / disk / bar decomposition is somewhat larger than in the decomposition where two bars were fitted ( b / t = 0.42 v.s .
0.34 ) , which means that in this case even the bar / oval ( visible as a bump at @xmath27 = 5 - 10 arcsec in the @xmath73-profile ) affects the solution .
0.25 cm * ngc 1350 * : this galaxy has been previously reported to have a prominent inner ring and an outer pseudo - ring of type r@xmath74 ( crocker , baugus @xmath5 buta 1996 ) . in the @xmath0-band ngc 1350
shows a prominent inner ring and an outer lindblad resonance subclass r@xmath74 pseudoring ( buta @xmath5 crocker 1991 ) . in the @xmath0-band ngc
1350 shows similar features .
the galaxy also shows a clear ansae - type bar , both in the direct images and in the @xmath19-profile .
the bar is more conspicuous at 2.2 @xmath73 m than in blue light but it is still relatively weak .
internal extinction masks much of the bar in blue light .
there is no clear sign of a secondary bar , but a peak appears in the @xmath19-profile at @xmath27 = 8 arcsec , associated with an isophotal twist , thus at least indicating that a separate structural component exists at that radius .
it might be an inner disk , because the orientation deviates only by 13@xmath22 from the orientation of the outer disk .
the bulge has a different orientation than the bar so that in our decomposition the two components can not be mixed .
0.25 cm the different components in the decomposition were found in small steps , and in the final solution the parameters of the disk and the length of the bar were fixed , the latter because for the ansae - type morphology the model would otherwise overestimate it .
again , in the robust 2d bulge / disk decomposition the flux of the bar goes erroneously to the bulge , giving @xmath4 = 0.77 , instead of @xmath4 = 0.25 , obtained in our best fitting solution . for this galaxy
the flattening of the bulge is high ( @xmath20 = 0.71 ) .
0.25 cm * ngc 1387 * : this galaxy is an excellent example of a very early - type barred galaxy .
although classified as type sab in rc3 , pahre ( 1999 ) and bco2007 adopted an sb type .
the @xmath75 family from paper ii is @xmath76b , indicating that the bar is indeed weak .
ngc 1387 has no spiral arms , but the surface brightness profile at @xmath27 = 30 - 90 arcsec is perfectly exponential , which is a strong sign of a disk . 0.25 cm in the decomposition we fitted , besides the bulge and the disk , also the bar / lens for which one single function was used .
subtraction of the bulge model from the original image showed that this galaxy has a prominent nuclear ring at @xmath27 = 6 arcsec ( fig .
5 ) , which appears also as a faint peak in the @xmath19-profile ( fig .
the ring is nearly circular and has a significant size compared to the length of the bar . in the surface brightness
profile the bulge and the disk appear as two nearly exponential functions .
0.25 cm * ngc 1411 * : sandage @xmath5 bedke ( 1994 ) noticed that ngc 1411 has a three - zone structure with a ring between the second and the third components .
bco2007 noted a series of lenses , and classified it as type sa(l)0@xmath22 , based on a @xmath66-band image . in our @xmath0-band image , only the innermost lens clearly stands out , but a weak outer lens , that has actually the appearance of a smooth ring , can also be detected in figure 5 , after subtracting the exponential disk .
the surface brightness profile of this galaxy shows three nearly exponential parts in the @xmath77-band image .
0.25 cm in spite of the simple looking profile , it was not easy to find a solution that converges . again the solution was found in small steps so that robust solutions for the two inner components were found first .
they were fixed and a solution for the outer disk was found .
after that the parameters of the outer disk were fixed and more precise solutions were found for the two inner components . finally all parameters of the bulge and the middle component were left free for fitting . the bulge and the middle component
were fitted using a sersic s function , whereas for the outer disk an exponential function was used .
our decomposition confirms the visual impression that all three components are nearly exponential .
we found @xmath4 = 0.08 , which is as small as the typical @xmath4-ratios in sc - type spirals .
our value for ngc 1411 is not essentially larger than the @xmath4 = 0.11 found by lsb2005 for sc - galaxies using the same decomposition method .
a similar small b / t ratio of 0.15 for sc - galaxies is given also by binney @xmath5 merrifield ( 1998 ) based on the @xmath72-band measurements by kent ( 1985 ) .
ngc 1411 might therefore be a genuine example of an s0c galaxy , that is , the proverbial analogue of an sc galaxy in a classical s0 , as discussed in the ddo galaxy classification scheme by van den bergh ( 1976 , 1998 ) .
0.25 cm a more robust 2d bulge / disk decomposition gives @xmath4 = 0.39 .
in this solution the @xmath4 value is very large , because the two inner exponentials are fitted by a single sersic s function .
for this galaxy the 1d decomposition ( see fig . 6 , lower left panel ) gives a very similar solution as the 2d bulge / disk decomposition ( middle lower panel ) . in its simplicity
ngc 1411 is an illustrative example : it clearly demonstrates that if a three - component galaxy is fitted by a two - component model , the two - dimensionality alone does not automatically improve the result .
0.25 cm * ngc 1512 * : this is a well - known interacting system and one of the latest hubble types in our sample . the morphology is dominated by a strong bar and two rings : a bright oval inner ring around of the bar ends and an intense nuclear ring in the center of the bar .
a deep @xmath78-band sings legacy image ( kennicutt et al .
2003 ) also reveals a distorted outer ring as well as faint tidal debris near the companion ngc 1510 .
the nuclear ring was first noted by hawarden et al .
( 1979 ) , who showed also the extent of ngc 1512 s interaction with neighboring ngc 1510 .
jungwiert , combes , @xmath5 axon ( 1997 ) noticed an isophotal twist at @xmath27 = 7 arcsec , which was suggested to be due to a triaxial bulge .
however , erwin ( 2004 ) used nicmos hst images to show that the twist is due to the nuclear ring , with no evidence of a bar inside the ring ( see also de grijs et al .
the ring ( with no bar inside ) is well visible also in our @xmath0-band image and in the @xmath15 color index map . as is typical of such rings , the morphology is patchy and indicative of strong star formation ( maoz et al .
0.25 cm as our image is not deep enough for a reliable decomposition for this galaxy , and the position angle is also very uncertain , we give only a rough estimate of the relative mass of the bulge .
only two components are used , a sersic s function for the bulge and an exponential function for the bar+disk .
the bulge has nearly circular isophotes and it is fitted by a function that approaches an exponential ( @xmath1 = 1.2 ) , thus indicating a disk - like nature of the bulge .
however , we stress that the decomposition is very uncertain , and therefore no @xmath4 flux ratio is given in table 4
. 0.25 cm * ngc 1533 * : bco2007 classify the galaxy as type ( rl)sb0@xmath22 , implying a large ring / lens outside the bar , but none of the conventional variety , either inner ring or inner lens .
sandage @xmath5 bedke ( 1994 ) noticed the slight enhancement at the rim of the rl and classified the galaxy as an sb0@xmath79/sba .
sandage @xmath5 brucato ( 1979 ) also noted the large lens under - filled by the bar , and they also suspected a weak spiral pattern in the outer lens . 0.25 cm for this galaxy it was possible to fit the bar and the lens separately . however , it was more difficult to find @xmath41 mainly because the lens dominates the disk in our image .
therefore @xmath41 was found using the 1d method by fitting an exponential to the outer disk . in the 2d decomposition @xmath41
was fixed to this value . * ngc 1553 * : the dominant feature of this galaxy s morphology is a strong inner ring / lens ( sandage @xmath5 brucato 1979 ; kormendy 1984 ) .
the lens shows subtle structure in our @xmath0-band image and is classified as an inner ring in rc3 . bco2007
give a type of sa(@xmath80l)0@xmath81 to highlight that the feature really is more a ring than a lens .
in the central part of the galaxy , a second elongated feature is present , which in our unsharp mask ( fig .
5 ) appears to be a nuclear disk showing also spiral arms .
sandage @xmath5 bedke ( 1994 ) noticed that this galaxy has subtle dust lanes in the disk leading to the classification of a peculiar sb0/sa .
the major - axis rotation curve for this galaxy is found to be typical for s0s ( kormendy 1984 ; rampazzo 1988 ) .
contrary to all previous evidence , sga2004 re - classified ngc 1553 as an elliptical with an inner disk , a choice which was justified by the fact that their decomposition did not find any outer disk .
no @xmath4-ratio is given by sga2004 , but the shape parameter of the bulge was found to be @xmath1 = 5.24 . 0.25 cm in our analysis the presence of the inner disk is obvious : it is particularly well visible in the image where the exponential disk is subtracted , shown in figure 5 .
the best 2d solution was found by fitting a bulge , lens / ring and an outer disk , which appeared to be exponential ( as also in the @xmath73-profile by kormendy 1984 ) .
we found @xmath4 = 0.21 and @xmath1 = 1.9 .
again , in order to duplicate the 2d solution of sga2004 we also made a 2d bulge / disk decomposition ; this gives @xmath4 = 0.88 and @xmath1 = 3.33 , which is exactly the same result as obtained also by our 1d method for this galaxy .
the reason for the very large difference between our 2- and 3-component decompositions is that in the 2-component solution the flux of the lens goes erroneously to the bulge . in all our decompositions
the shape parameter of the bulge is considerably smaller than the value @xmath1 = 5.24 obtained by sga2004 .
0.25 cm the nature of the bulge in this galaxy is unclear even in our best fitting solution .
the component that is fitted as a bulge is actually the inner disk and there is no other more spherical component in the central regions that could be interpreted as a bulge .
perhaps there is no bulge in this galaxy , unless the inner disk is interpreted as a pseudo - bulge .
kinematical evidence for this interpretation comes from the study by kormendy ( 1984 ) who argued that the bulge ( the small central component ) is cooler than the lens dominating at larger radial distances .
the change in the stellar velocity dispersion occurs at @xmath27 = 12 -15 arcsec after which the lens starts to dominate .
this is precisely the distance inside of which we find the disk - like bulge in our decomposition .
0.25 cm * ngc 1574 * : bco2007 classify this galaxy as type sb0@xmath8 , where the bar is a well - defined feature embedded in a huge extended disk component .
the bar was also noted by sandage @xmath5 bedke ( 1994 ) but was missed in rc3 . de vaucouleurs @xmath5 de vaucouleurs ( 1964 ) pointed out the presence of a lens . although sandage @xmath5 bedke note that ngc 1574 has a three - zone structure typical of s0s , bco2007 could not identify a clear inner lens around the bar . from our @xmath0-band image
it is difficult to judge for sure whether this galaxy has a lens or not , but the surface brightness profile shows three nearly exponential parts , which is the case also in some other galaxies with lenses in our sample .
the bar is embedded inside the middle exponential component of the disk .
0.25 cm the decomposition was carried out starting from the parameters of the outer disk . once reasonable approximations were found for all components ,
better solutions were searched separately for each of them . finally all parameters describing these components ( except for the ellipticity of the lens ) were left free for fitting .
contrary to the usual case , for this galaxy the decomposition depends critically on the image resolution : by rebinning the image by a factor of 4 we find @xmath4 = 0.28 and @xmath1 = 1.5 , whereas for the original image we obtain @xmath4 = 0.38 and @xmath1 = 2.9 .
a possible reason for the large difference between these two solutions is the compact central cusp at @xmath27 = 9 arcsec , found by phillips et al .
( 1996 ) : most probably the cusp is better resolved in the original image .
the fitted lens inside the bar region mainly improves the model image , but does not affect the properties of the bulge . subtracting the bulge model from the original image
shows a complete inner ring surrounding the bar , which changes the rc3 s - variety classification to r - variety .
0.25 cm * ngc 3081 * : ngc 3081 is a well - known seyfert 2 galaxy having four exceptionally well - defined rings ( buta @xmath5 purcell 1998 ) : an r@xmath69 outer ring intrinsically aligned nearly perpendicular to the bar , a nearly circular r@xmath70 outer pseudoring , a highly elongated , aligned inner ring , and a misaligned nuclear ring .
the deprojected radii of these features are 90 , 105 , 35 , and 14 arcsec , respectively , and color index maps ( buta 1990 ; buta @xmath5 purcell 1998 ) indicate that each ring is a well - defined zone of active star formation .
the strong similarity between these rings and test - particle models of barred spirals ( simkin , su , @xmath5 schwarz 1980 ; schwarz 1981 ) led buta @xmath5 purcell ( 1998 ) to classify ngc 3081 as a resonance ring galaxy .
the galaxy also has a complex double - barred structure detected both in the optical and in the near - infrared ( buta 1990 ; friedli et al .
1996 ; mulchaey , regan @xmath5 kundu 1997 ; erwin 2004 ; laine et al .
the primary bar is a weak feature that underfills the bright inner ring .
the secondary bar lies within the nuclear ring and is misaligned with the primary bar by 71@xmath22 . except for the r@xmath70 outer pseudoring ,
all of these structural components of ngc 3081 are visible in our @xmath0-band image .
the secondary bar is also surrounded by a lens extending to the radius of the nuclear ring . as noted in paper ii
, the main bar - like feature in ngc 3081 is the massive oval rimmed by the inner ring .
the apparent primary bar underfills this ring / oval and seems insignificant in comparison . in spite of its early morphological type ( s0/a ) , ngc 3081 has @xmath82 , @xmath83 , and @xmath84 colors similar to those typically found in spiral galaxies ( alonso - herrero et al .
the hydrogen index also indicates a higher than average hi content for its type ( buta 1990 ) .
star formation in the nested rings has also been studied by ferruit , wilson @xmath5 mulchaey ( 2000 ) , buta , byrd @xmath5 freeman ( 2004 ) , and martini et al .
( 2003 ) although in the circumnuclear regions ( 210 - 590 pc ) the stellar population is largely dominated by old stars ( boisson et al .
buta , byrd , @xmath5 freeman ( 2004 ) show high resolution hst images of the inner regions of ngc 3081 , revealing hundreds of barely - resolved star clusters in the inner ring .
the distribution of these clusters belies the intrinsic oval shape of the ring , because they concentrate in arcs around the ring major axis .
evans et al .
( 1996 ) also noted that ngc 3081 has a bright nucleus with an apparent bulge visible to the radius of @xmath27 = 10 arcsec . 0.25 cm in the decomposition we fitted the bulge , disk , the secondary bar / lens and the primary bar .
it was difficult to find a solution that converges , which is not surprising taking into account the rich morphology of this galaxy .
as always in complicated cases , a reasonable solution was found by repeating a series of decompositions . in the inner regions it was not possible to fit simultaneously the bulge , lens and the secondary bar , mainly because of the presence of the strong nuclear ring . instead , a solution was found were the nuclear bar and the lens were fitted by a single function .
also , a simultaneous presence of a weak primary bar and a strong ring leads to a solution where the bar fills the region up to the radius of the inner ring .
this somewhat overestimates the flux of the primary bar .
however , this is not expected to seriously underestimate the flux of the bulge , which resides in the central regions of this galaxy . in the residual image
both the inner and the nuclear ring are clearly visible , and the secondary bar shows an ansae - type morphology . 0.25 cm * ngc 3100 * : in the early work by sandage @xmath5 brucato ( 1979 ) weak spiral arms were detected in the outer disk , although in rc3 this galaxy is classified as a weakly barred peculiar s0 . in our analysis
ngc 3100 appears as a system with a large lens and a bright nuclear component , to which already de vaucouleurs @xmath5 de vaucouleurs payed attention in 1964 . evidently , there are also weak asymmetric spiral arms in the outer disk , as demonstrated by the image where the disk model is subtracted ( fig .
the bulge component is difficult to evaluate in this galaxy
. it could be the bright nuclear component or there is no bulge at all .
the larger component fitted by a sersic s function is clearly the large lens having @xmath1-parameter ( @xmath1 = 1.8 ) typical for disky systems .
as we were not able to fit the small nuclear component , no @xmath4-ratio is given in table 6 .
this is our second candidate of galaxies that might be a proverbial analogue of late - type spirals among s0s . 0.25 cm * ngc 3358 * : the surface brightness of this early - type spiral ( bco type ( r@xmath70)sab(l)a ) is high and the main morphological features lie on an extended , almost flat distribution of light .
the galaxy is most noteworthy for its strong r@xmath70 outer pseudoring which is made from two bright arms that begin well off the ends of the bar and which partly overshoot those ends .
the overshoot indicates a trace of an r@xmath69 component . in optical images ,
the outer pseudoring is lightly patchy and is affected by dust lanes on the near side .
the bar is a weak ansae type , but having also some characteristics of spiral arms ( see figure 5 ) .
the bar is embedded in a well - defined inner lens , which causes conflicting interpretations of the variety .
there is trace of a secondary bar in our @xmath0-band image , previously detected in the optical region by buta @xmath5 crocker ( 1993 ) ( see also erwin 2004 ) .
the bulge seems to be the very small , nearly spherical component inside the secondary bar .
unfortunately , the structure of this galaxy was too complicated for any reasonable decomposition .
because of the freeman type ii surface brightness profile , not even a 1d decomposition was made .
0.25 cm * ngc 3706 * : this galaxy seems to have a very small , nearly edge - on inner disk , embedded inside a prominent bulge .
the inner disk has been directly detected in the hst image ( filter = f555w ) by lauer et al.(2005 ) being visible above the bulge at least to @xmath27 = 1 arcsec .
they found also a stellar ring at that radius .
it is possible to fit this galaxy either by a single de vaucouleurs type function , or by adding an exponential inner disk .
the two - component solution gives a marginally better fit and above all , it represents better the pre - identified structural components of this galaxy . in that case also
the shape parameter of the bulge is more reasonable : @xmath1 = 3.6 instead of @xmath1 = 5.1 .
our best solution is very similar to that found by scorza et al .
( 1998 ) in the @xmath66-band : in both cases @xmath4 = 0.92 was found ( scorza et al .
1998 gives @xmath67 = 0.06 ) .
as in our @xmath0-band image a subtle trace of an inner lens is visible , we interpret the galaxy as type e@xmath74 .
it is generally assumed that bars in 2d decompositions do not affect the relative masses of bulges or the properties of bulges and disks , an understanding which is largely based on the study by de jong ( 1996 ) , who compared 2d bulge / disk and bulge / disk / bar decompositions for spirals .
later it was shown by peng et al .
( 2002 ) for a few individual galaxies , and by lsb2005 for a larger number of early - type disk galaxies , that bars might actually play an important role in the structural decompositions .
these studies applied more advanced 2d methods than used in the study by de jong .
in particular it was shown by lsb2005 that the omission of a bar might overestimate the mass of the bulge even by 40@xmath85 .
another generally accepted assumption is that 2d methods always give more reliable results than 1d methods . in the following
we evaluate the reliability of this assumption .
0.25 cm the mean values for the parameters of the bulge and the disk were calculated by including in the statistics all those 13 galaxies in our sample for which it was possible to estimate the @xmath4-ratio in a reasonable manner ( excluded are ngc 3358 and ngc 3100 ) , as well as the 11 disk - dominated systems .
the results are shown in table 7 .
a comparison of 2d 2- and 3-component decompositions shows that when the primary bars or ovals are omitted , the @xmath4 ratios are significantly overestimated , mainly because the flux of the third component goes erroneously to the bulge . in the 3-component decompositions the bulges are also more exponential than while using the 2-component method .
however , including more components to the fit does not affect the mean @xmath4-ratio considerably .
multiple bars / ovals affect the @xmath4-ratio only in galaxies with very small bulges or having high surface brightness secondary bars or ovals .
our best fitting multicomponent method gives @xmath86 = 0.25 and @xmath87 = 2.1 , which are very similar to @xmath86 = 0.24 and @xmath87 = 2.1 as found earlier by lsb2005 for a sample of 24 northern early - type disk galaxies .
this @xmath4 flux ratio is considerably smaller than obtained by our 2d bulge / disk approach , which gives @xmath86 = 0.55 . or if the two elliptical galaxies classified as s0s by sga2004 are also included , @xmath86= 0.61 .
this is very similar as the mean @xmath4 ratio of 0.61 obtained by sga2004 using a similar 2d bulge / disk decomposition approach in the same wavelength .
0.25 cm we also find that @xmath41 is not sensitive to the 2d method used , indicating that a bar or an oval does not affect the scale length of the disk .
the 1d method generally also gives similar @xmath41 , but there are some individual cases where different @xmath41 values are obtained .
the 1d method failed to find the extended disks for the galaxies ngc 1079 ( @xmath41 = 14 arcsec v.s .
34 arcsec ) and ngc 1350 ( @xmath41=47 arcsec v.s .
141 arcsec ) . for ngc 3081
the 1d method gives larger @xmath41 ( 38 arcsec v.s 21 arcsec ) , mainly because the 1d fit extends to a larger radial distance .
therefore in this case the 1d method gives a more reliable estimate of @xmath41 .
it is due to these exceptional cases that makes the small difference in the mean @xmath41 obtained by the 1d and 2d methods in table 7 . in conclusion
, we did not find any evidence supporting the view that for multicomponent galaxies the 2d method automatically gives better results than the 1d method .
this is the case only if , in addition to the bulge and the disk , at least one more component is included to the 2d fit
for any hypothesis of galaxy evolution it is important to understand whether bulges in galaxies are old classical bulges , or pseudo - bulges formed at later phases of galaxy evolution .
the main approaches for evaluating this are the kinematic and photometric approaches .
bulges are expected to have the characteristics of pseudo - bulges mainly when @xmath88 2 ( kk2004 ) .
this might be the case also when the bulge is oriented near to the plane of the outer disk , or the ellipticity of the bulge is near to that of the outer disk .
however , the bulge is not necessarily aligned with the disk .
s0s might have many disks with different orientations , of which ngc 1411 is a good example .
it is also assumed by kk2004 that if a galaxy has a secondary bar or a nuclear ring inside the bulge , the bulge is most probably a young pseudo - bulge .
even more direct evidence for the existence of a pseudo - bulge can be obtained kinematically : if the ratio of the maximum velocity amplitude ( v@xmath89 ) to the mean stellar velocity dispersion ( @xmath90 ) at the same radius , v@xmath91 , versus the ellipticity ( @xmath19 ) is above the isotropic oblate rotator curve @xcite , the bulge is kinematically supported .
this means that stars of the bulge have not had time to gain such large random velocities as in the case of older classical bulges .
unfortunately useful kinematic data is available only for a few galaxies in our sample .
0.25 cm we found strong evidence of pseudo - bulges in the non - elliptical galaxies in our sample .
the shape parameter of the bulge is smaller than 2 in six of the 12 galaxies , and near to @xmath1=2 in ten of the galaxies . for ngc 3358
no decomposition could be made , but as discussed above , the bulge also for this galaxy is very small , most probably having the characteristics of pseudo - bulges .
larger @xmath1-values for the bulge were detected only for ngc 1326 ( @xmath1 = 3.0 ) and ngc 1574 ( @xmath1 = 2.9 ) , but even ngc 1326 has a nuclear star forming ring , and for ngc 1574 there is strong kinematic evidence showing that the bulge is rotationally supported . for ngc 1574
we use @xmath90 and v@xmath89 measured by jarvis et al.(1988 ) .
they found v@xmath89(obs ) = 40 km sec@xmath92 along the major axis , and after correcting the inclination effect this is converted to v@xmath89(corr ) = 145 km sec@xmath92 .
based on our surface brightness profile , the galaxy is bulge dominated at r = 6 arcsec , at which radius @xmath90 = 180 km sec@xmath92 and @xmath19 = 0.1 ( taken from our @xmath19-profile in fig .
the obtained v@xmath89/@xmath90 = 0.8 , which at @xmath19 = 0.1 means that the bulge is rotationally supported . in a similar manner v@xmath89/@xmath90
were found for the galaxies ngc 1553 , ngc 3100 , ngc 3706 and es0 208-g21 . for ngc 3100 we used v@xmath89(corr ) = 120
km sec@xmath92 and @xmath90 = 190 km sec@xmath92 ( carollo , danziger @xmath5 buson 1993 ) .
the measurements were made at the position angle of 90@xmath22 , which deviates only 10@xmath22 from the orientation of the major axis of the disk . in the bulge - dominated region at @xmath27 = 5 - 10 arcsec this means that the bulge is rotationally supported . for ngc 1553
we use v@xmath89(corr ) = 156 km sec@xmath92 and @xmath90 = 162 km sec@xmath92 at @xmath27 = 10 arcsec ( kormendy 1984 ) , which also in this case means that the bulge is rotationally supported .
quite interestingly , according to kormendy the bright lens dominating at larger radial distances is hotter and not rotationally supported , indicating that _ the lens is probably older than the bulge_. as expected , the elliptical galaxies ngc 3706 and eso 208-g21 have neither kinematic nor photometric evidence of pseudo - bulges .
four of the galaxies in our sample have also either a nuclear ring , a secondary bar or both , and generally the nuclear ring is presently forming stars . in conclusion , all galaxies with extended disks in our sample have some sign of a pseudo - bulge .
pseudo - bulges have been previously found in many late - type spirals ( carollo et al . 1998 ) , and more recently also in two early - type spirals ( kormendy et al .
2006 ) with the hubble types sab and sa .
0.25 cm an other interesting new finding of this study is that the bulges of two s0s in our sample are as small as in typical sc - type spirals . for ngc 1411 @xmath4=0.08 in comparison to
0.11 found in sc - galaxies .
for ngc 3100 the bulge is also very small , but could not be fitted in our decompositions .
both galaxies are dominated by lenses , ngc 1411 by a series of bright lenses , and ngc 3100 by one dominant lens / oval .
the bulge in ngc 1411 is disk - like ( have small @xmath1 ) , but none of the two galaxies have a nuclear ring , or other strong morphological signatures of strong star formation in the central regions of these galaxies .
the bulge in ngc 1553 most probably is also extremely small , unless the inner disk ( with obvious spiral arms ) is considered as a pseudo - bulge . to our knowledge
ngc 1411 and ngc 3100 are the first candidates that might be analogs of sc - galaxies among s0s , with similar @xmath4- flux ratios as typically found in sc - galaxies .
the existence of such galaxies has been earlier suggested in the ddo classification scheme by van den bergh ( 1976 , 1998 ) .
these two galaxies are extreme examples of s0s that would be difficult to explain by merging of large disk galaxies .
ngc 1411 is found to be a member of the dorado group of 46 galaxies ( sandage 1975 ; maia , da costa @xmath5 latham 1989 ) , which is one of the richest galaxy groups in the southern hemisphere . using a more strict definition of a group , garcia ( 1993 ) found the same galaxy to be a member of a small group of galaxies .
ngc 1411 has also one major nearby companion , whereas ngc 3100 has many small nearby companions . particularly in case of ngc 1411
it is possible that a late - type spiral has transformed to an s0c type system by stripping mechanisms .
0.25 cm ovals or lenses were found in 11 of the galaxies in our sample .
ovals have minor to major axis ratios between 0.85 and 0.89 , in agreement with the previous estimations by kormendy ( 1984 ) for typical ovals .
however , the presence of lenses is more difficult to show , because their isophotes are not necessarily elongated .
this is also the reason why they are not always detected while classifying galaxies . in the decompositions made in this study
, the lenses can be best recognized due to the model images ( produced in the structural decompositions ) that should have the observed galaxy morphologies . in many cases a good agreement between the model image and the observation was found only if a flat inner structure was also added to the fit .
galaxies with prominent lenses in our sample typically have multiple exponential parts in their surface brightness profiles .
0.25 cm bars , bulges and halos are related phenomena in galaxies .
it has been shown by the theoretical models ( see the review by athanassoula 2005 ) that the properties of bars are formed in a complicated process between the halo and the disk , for example depending on the mass of the halo , the amount of halo mass in the resonances , or the mass distribution of the bulge .
when these properties are favorable to angular momentum exchange between the halo and the disk , bars are expected to evolve over time , leading to more massive bars with larger quadrupole moments . as bars loose angular momentum in this process
these models also predict that bars gradually slow down .
however , the models are still controversial . while the bar appears to slow down in most n - body models ( hernquist @xmath5 weinberg 1992 ; debattista @xmath5 sellwood 2000 ; athanassoula 2003 ; sellwood 2003 ; holley - bockelmann , weinberg @xmath5 katz 2005 ) , there are some other models in which the bar slow down is barely visible over the hubble time ( valenzuela @xmath5 klypin 2003 ) .
also , because long bars in early - type disk galaxies are repeatedly found to be fast rotating systems ( see the review by debattista 2006 ) it is not clear that the longest bars are necessarily the most evolved bars . in this study
it is interesting to look at whether the properties of bars in s0s are similar or fundamentally different from the properties of bars in early - type spirals . 0.25 cm in table 8 we compare the mean properties of primary bars in different hubble type bins . the properties for sbc - sm ( t = 4 - 9 ) type galaxies are from lsb2004 , based on a similar analysis for the galaxies in the osubg - survey . in order to increase the statistics for the sa - sab type spirals ,
the data is a collection from three samples : the osubgs , the sample of early - type disk galaxies by lsb2005 , and this study .
there is a clear tendency showing that bar length increases from the late - type spirals towards the early - type spirals , as first suggested by elmegreen @xmath5 elmegreen ( 1985 ) .
the maximum is reached for the hubble type s0/a , but bars in s0s are clearly shorter than bars in s0/a galaxies , although still longer than bars in late - type spirals .
this is in agreement with the recent result obtained by erwin ( 2005 ) .
the relative mass of the bar , as estimated from the @xmath32 and @xmath33 fourier amplitudes of density , has a similar tendency .
however , the bar induced maximum tangential force , @xmath28 , decreases all the way from late - type spirals towards earlier types , reaching a minimum for s0s .
the fact that @xmath28 is slightly smaller in s0s is due more to the smaller bar size and mass than to dilution by the bulge , because the b / t ratios in s0s and s0/a galaxies are fairly similar . the way how massive bulges dilute @xmath28 was shown in lsb2004 .
0.25 cm an interesting result in this study is that the properties of bars in the spiral / s0 divide , e.g. between the s0/a and s0-galaxies are not exactly similar , or explainable by the dilution effects of bulges .
bars in s0s are shorter ( 1.21 v.s . 1.72 , respectively ) and less massive ( @xmath32 = 0.49 v.s .
0.63 ) than bars in s0/a - galaxies .
the fact that @xmath28 is slightly smaller in s0s is associated to the mass of the bar rather than to the dilution by bulge , because the b / t - ratios is s0s and s0/a - galaxies are fairly similar ( @xmath28= 0.24 @xmath930.14 v.s .
0.27 @xmath93 0.17 ) .
the weaker bars in s0s might be due to a smaller amount of interstellar gas in these galaxies , which is expected to make the disks less reactive to bar formation .
bars in s0s might be shorter and less massive also because they are less evolved in time ( see the models by athanassoula 2003 ) . 0.25 cm
a morphological analysis of a sample of 15 early - type disk galaxies , mainly s0s , has been presented in the @xmath0-band .
the sample is part of a more complete nirs0s - sample of 170 s0-s0/a galaxies , selected to have a similar size and selection criteria as in the osubgs - sample for spirals .
our main idea is to characterize the properties of bars , ovals and bulges for a large sample of galaxies in the spiral / s0 divide and to use this information to evaluate the origin of s0s and how they are related to spirals .
0.25 cm the galaxies were analyzed by applying fourier methods for calculating the maximum gravitational torques , and by applying a multicomponent 2d method for structural decompositions .
a generalized sersic s function is used for the bulge and exponential function for the disk , whereas the other non - axisymmetric components can be described either by a sersic s or a ferrers function .
it was also possible to take into account the flattening and the shape parameter of the structures . for comparison
1d decomposition method was also used .
we found that if only the bulge and the disk are fitted , 1d and 2d methods give fairly similar results ( @xmath86 = 0.55 - 0.61 ) .
0.25 cm bars in the galaxies in our sample are typically surrounded by lenses , and four of the galaxies have double bar / lens structures .
two of the galaxies classified as s0s in rc3 appeared to be elliptical galaxies ( eso 208-g21 , ngc 3706 ) , and three of the non - barred s0s were found to have inner disks ( ngc 1411 , ngc 3100 , ngc 1553 ) .
a previously non - detected nuclear ring was discovered in ngc 1387 , and a nuclear bar in ngc 1326 .
we found that bars in s0s are not exactly similar to bars in s0/a - galaxies .
they are clearly shorter and less massive and have slightly weaker bar torques than bars in s0/a - galaxies .
0.25 cm one of the most important results of this study is the finding that _ 11 non - elliptical galaxies in our sample show strong evidence of having disk - like pseudo - bulges _ : these bulges have surface brightness profiles that are more close to an exponential than to a de vaucouleurs type profile .
six of the galaxies also have either nuclear rings , nuclear bars or nuclear disks .
all galaxies in our sample have @xmath94 0.4 as typically found in galaxies having pseudo - bulges .
we find @xmath86 = 0.25 and @xmath87 = 2 , thus confirming the earlier result by lsb2005 , showing that bulges in s0s are considerably smaller and have more exponential surface brightness profiles than assumed for their morphological types .
a comparison to more simple 2d bulge / disk or 1d - decompositions shows the importance of the multicomponent approach : these two methods give @xmath86 = 0.55 - 0.48 , respectively , in both cases the relative mass of the bulge being significantly overestimated .
the value @xmath86 = 0.55 is close to the value 0.64 obtained by sga2004 for the early - type disk galaxies in the same wavelength , using a similar 2d bulge / disk approach as used in this study
. 0.25 cm quite interestingly , we also find two galaxies , ngc 1411 and ngc 3100 , which due to their very small @xmath4 flux ratios might be proverbial analogues of sc galaxies among classical s0s , as discussed in the ddo galaxy classification scheme by van den bergh .
these galaxies might be manifestations of galaxy evolution in the hubble sequence , possibly formed by gas stripping mechanisms from late - type spirals .
el and hs acknowledge the support of the academy of finland while preparing this manuscript .
rb also acknowledges the support of nsf grant ast0507140 to the university of alabama . this research has also made use of the nasa / ipac extragalactic database ( ned ) which is operated by the jet propulsion laboratory , california institute of technology , under contract with the national aeronautics and space administration .
jhk acknowledges support from the leverhulme trust in the form of a leverhulme research fellowship .
we also acknowledge eso to make possible for carrying out this observing run .
alonso - herrero , a. , simpson , c. , ward , m. , wilson , a. 1998 , , 495 , 196 athanassoula , e. 2003 , , 341 , 1178 athanassoula , e. , morin , s. , wozniak , h. , puy , d. , pierre , h. , lombard , j. , bosma , a. 1990 , , 245 , 130 athanassoula , e. 2005 , celestial mechanics and dynamical astronomy 91 , 9 - 31 ( review ) bekki , k. , warrick , j. , yasuhiro , s. 2002 , , 577 , 651 binney , j. 1980 , , 190 , 421 binney , j. , merrifield , m. 1998 , `` galactic astronomy '' ( princeton university press ) buta , r. 1988 , , 66 , 233 buta , r. 1990 , , 351 , 62 buta , r. 1995 , , 96 , 39 buta , r. , crocker , d. 1991 , , 192 , 1715 buta , r. , crocker , d. 1993 , , 105 , 1344 buta , r. , purcell , g. 1998 , , 115 , 484 buta , r. , alpert , a. , cobb , m. , crocker , d. , purcell , g. 1998 , , 116 , 1142 buta , r. , treuthardt , p. , byrd , g. , crocker , d. 2000 , , 121 , 255 buta , r. , byrd , g. , freeman , t. 2004 , , 127 , 1982 buta , r. , block , d. 2001 , , 550 , 243 buta , r. , laurikainen , e. , salo , h. , block , d. , knapen , j. , 2006 , aj , in press ( paper ii ) buta , r. , corwin , h. , odewahn , s. 2006 , `` the de vaucouleurs atlas of galaxies '' , cambridge , cambridge university press ( bco2007 ) boisson , c. , joly , m. , pelat , d. , ward , m. 2004 , 428 , 373 carollo , c. , stiavelli , m. , mack , j. 1998 , , 116 , 68 carollo , c. , danziger , i. , buson , l. 1993 , , 265 , 553 christlein , d. , zabludoff , a. 2004 , , 616 , 192 ( cz04 ) combes , f. , sanders , r. h. 1981 , , 96 , 164 corwin , h. g. , de vaucouleurs , a. , de vaucouleurs , g. in southern galaxy catalogue of 5481 galaxies south of declination 17 degrees found in 1.2 m u.k .
smidth iiia - j plates , austin , texas , 185 crocker , d. , baugus , p. , buta , r. 1996 , 105 , 353 debattista , v. 2006 , astro - ph/0601277 debattista , v. , sellwood j.a .
2000 , , 543 , 704 de grijs , r. , anders , p. , bastian , n. , lynds , r. , lamers , h. , oneil , e. 2003 , , 343 , 1285 de jong , r. , 1996 , , 313 , 45 de vaucouleurs , g. , de vaucouleurs , a. 1964 , `` reference catalogue of bright galaxies '' ( austin : university of texas ) de vaucouleurs , g. 1974 , in formation and dynamics of galaxies , iau symposium no .
58 , j. r. shakeshaft , ed . , dordrecht , reidel , p. 1 .
de vaucouleurs , g. , de vaucouleurs , a. , corwin , h. , buta , r. j. , paturel , g. , fouqu , p. 1991
, `` third reference catalog of bright galaxies '' ( new york : springer ) ( rc3 ) de souza , r. , gadotti , a. , dos anjos , s. 2004 , , 153 , 411 ( sga2004 ) erwin , p. 2004
, , 415 , 941 erwin , p. 2005
, , 364 , 283 eskridge , p. et al .
2002 , , 143 , 73 evans , i. , koratkar , a. , storchi - bergmann , t. , kirkpatrick , h. , heckman , t , wilson , a. 1996 , , 105 , 93 ferruit , p. , wilson , a. , mulchaey , j. 2000 , , 128 , 139 fernandes , c. , storchi - bergmann , t. , schmitt , h. 1998 , , 297 , 579 friedli , d. , wozniak , h. , rieke , m. , martinet , l. , bratschi , p. 1996
, , 118 , 461 garcia , a. m. 1993 , , 100 , 47 garcia - barreto , j. , dettmar , r. , combes , f. , gerin , m. , koribalski , b. 1991 , , 22 , 197 garcia - gomez , c. , barbera , c. , athanassoula , e. , bosma , a. , whyte , l. 2004 , , 421 , 595 hernquist , l. , weinberg , m. 1992 , , 446 , 717 hawarden , t. , van woerden , h. , goss , w. , mebold , u. , peterson , b. 1979 , , 76 , 230 holley - bockelmann , k. , weinberg , m. , katz , n. 2005 , , 363 , 991 jarvis , b. , dubath , p. , martinet , l. , bacon , r. 1988 , , 74 , 513 jungwiert , b. , combes , f. , axon , d. 1997 , , 125 , 479 kennicutt , r. c. jr .
2003 , , 115 , 928 kent , s. 1985 , , 59 , 115 knapen , j. h. , mazzuca , l. m. , bker , t. , shlosman , i. , colina , l. , combes , f. , & axon , d. j. 2006 , , 448 , 489 kormendy , j. , 1979 , , 227 , 714 kormendy , j. , 1984 , , 286 , 116 kormendy , j. , bender , r. 1996 , , l119 kormendy , j. , kennicutt , r. jr . 2004 , ann rev .
42 , 603 ( kk2004 ) 111 kormendy , j , cornell , , m. e. , block , d. , knapen , j. , allard , e. 2006 , , 642 , 765 kuchinski , l. , freedman , w. , madore , b. , trewhella , m. , bohlin , r. , cornett , r. , fanelli , m. , marcum , p. , neff , s. , oconnell , r. and 4 coauthors 2000 , , 131 , 441 laine , s. , shlosman , i. , knapen , j. h. , & peletier , r. f. 2002 , , 567 , 97 laurikainen , e. , salo , h. 2000 , , 141 , 103 laurikainen , e. , salo , h. , 2002 , 337 , 1118 laurikainen , e. , salo , h. , buta , r. , vasylyev s. , 2004 , , 355 , 1251 laurikainen , e. , salo , h. , buta , r. , 2004 , , 607 , 103 ( lsb2004 ) laurikainen , e. , salo , h. , buta , r. 2005 , , 362 , 1319 ( lsb2005 , paper i ) lauer , t. , faber , s. , gebhardt , k. , richstone , d. , tremaine , s. , ajhar , e. , aller , m. , bender , r. , dressler , a. , filippenko , a. , and 7 coauthors 2005 , , 129 , 2138 maia , m. a. , da costa , l. n. , latham , d. w. 1989 , , 69 , 809 martel , a. , ford , h. , bradley , l. , tran , h. , menanteau , f. , tsvetanov , z. , illingworth , g. , hartig , g. , clampin , m. 2004 , , 128 , 2758 maoz , d. , barth , a. , ho , l. , sternberg , a. , filippenko , a. 2001 , 121 , 3048 martini , p. , regan , m. , mulchaey , j. , pogge , r. 2003 , , 146 , 353 mulchaey , j. , regan , m. , kundu , a. 1997 , , 110 , 299 pahre , m. 1999 , , 124 , 127 papovich , c. , giavalisco , m. , dickinson , m. , conselice , c. , ferguson , h. 2003 , , 598 , 827 peng , c. , ho , l. , impey , c. , rix , h. 2002 , , 124 , 266 pfenniger , d. 1984 , , 134 , 373 phillips , a. , illingworth , g. , mackenty , j. , franx , m. 1996 , , 111 , 1566 quillen , a. , frogel , j. , gonzalez , r. 1994 , , 437 , 162 raha , n. , sellwood , j. , james , r. , kahn , f. 1991 , nature , 352 , 411 rampazzo , r. 1988 , , 204 , 81 salo , h. , rautiainen , p. , buta , r. , purcell , g. , cobb , m. , crocker , d , laurikainen , e. 1999 , , 117 , 792 salo , h. , laurikainen , e. , buta , r. 2004 , in `` penetrating bars through masks of cosmic dust '' , eds .
d. block , i. puerari , k. freeman , r. groess , e. block , ( springer ) , p. 673 sandage , a. 1975 , , 202 , 56 sandage , a. , bedke j. 1994 , `` the carnegie atlas of galaxies '' ( washington : carnegie inst.)(cag ) sandage , a. , brucato r. 1979 , , 84 , 472 schweizer , f. 1980 , , 237 , 303 schwarz , m. 1981 , , 247 , 77 scorza , c. , bender , r. , winkelmann , c. , capaccioli , m. , macchetto , d. f. 1998 , , 131 , 265 sellwood , j.a .
2003 , , 587 , 638 simkin , s. , su , h. , schwarz , m. 1980 , , 237 , 404 storchi - bergmann , t. , rodriguez - ardila , a. schmitt , h. , wilson , a. , baldwin , j. 1996 , , 472 , 83 valenzuela , o. , klypin , a. 2003 , , 345 , 406 van den bergh , s. 1979 , , 206 , 883 van den bergh , s. 1998 , `` galaxy morphology and classification '' ( cambridge university press ) wozniak , h. , friedli , d. , martinet , p. , bratschi , p. 1995
, , 111 , 115 lllccc & & & & + galaxy & @xmath95 & @xmath96 & filter & fwhm & stdev + & & & & ( arcsec ) + eso 208@xmath97g21 & e(d)5 & sab@xmath8 & @xmath0 & 0.63 & 1.1 + ngc 1079 & ( r@xmath69r@xmath70)s@xmath76b(@xmath80s)a & ( r)sab(rs)0/a & @xmath0 & 0.84 & 1.0 + ngc 1317 & ( r@xmath98)sab(rl)a & sab(r)a & @xmath14 & 0.60 & 1.0 + & & & @xmath0 & 0.66 & 2.4 + ngc 1326 & ( r@xmath69)sab(r)0/a & ( r)sb(r)0@xmath7 & @xmath14 & 0.72 & 1.0 + & & & @xmath0 & 0.72 & 1.0 + ngc 1350 & ( r@xmath74)sab(r)ab & ( r@xmath99)sb(r)ab & @xmath0 & 0.75 & 1.5 + ngc 1387 & sb0@xmath8 & sab(s)0@xmath8 & @xmath0 & 0.69 & 1.1 + ngc 1411 & sa(l)0@xmath22 & sa(r)0@xmath8 & @xmath0 & 0.81 & 1.2 + ngc 1512 & r@xmath98)sb(r)ab pec & sb(r)a & @xmath14 & 0.69 & 2.1 + & & & @xmath0 & 0.63 & 1.0 + ngc 1533 & ( rl)sb0@xmath22 & sb0@xmath8 & @xmath0 & 0.66 & 1.3 + ngc 1553 & sa(@xmath80l)0@xmath81 & sa(rl)0@xmath22 & @xmath0 & 0.75 & 1.2 + ngc 1574 & sb0@xmath8 & sa(s)0@xmath8 : & @xmath0 & 0.75 & 1.3 + ngc 3081 & ( r@xmath69r@xmath70)sab(r)0/a & ( r)sab(r)0/a & @xmath14 & 0.69 & 1.5 + & & & @xmath0 & 0.63 & 2.0 + ngc 3100 & sab(s)0@xmath7 pec & sab(s)0@xmath22 pec & @xmath14 & 0.84 & 2.5 + ngc 3358 & ( r@xmath70)sab(l)a & ( r)sab(s)0/a & @xmath0 & 0.78 & 0.9 + ngc 3706 & e@xmath74 & sa(rs)0@xmath8 & @xmath0 & 0.81 & 0.8 + ccllcrrl galaxy & filter & @xmath18 ( measured ) & @xmath20 ( measured ) & range & @xmath18 ( rc3 ) & q ( rc3 ) & @xmath100 ( other ) + & & ( degrees ) & & ( arcsec ) & ( degrees ) & & ( degrees)/ + eso 208-g21 & @xmath0 & * 109 @xmath93 1 * & * 0.722 @xmath93 0.024 * & 93 - 116 & 110 & 0.708 & 108/0.70 + ngc 1079 & @xmath0 & * 78 @xmath93 2 * & * 0.590 @xmath93 0.011 * & 116 - 128 & 87 & 0.607 & + ngc 1317 & @xmath14 & * 91 @xmath93 2 * & * 0.953 @xmath93 0.004 * & 75 - 87 & 78 & 0.871 & + ngc 1326 & @xmath14 & 75.6@xmath93 1.1 & 0.729 @xmath93 0.000 & 134 - 140 & 77 & 0.741 & * 73.4/0.750 * + ngc 1350 & @xmath0 & * 6 @xmath93 1 * & * 0.586 @xmath93 0.004 * & 120 - 125 & 0 & 0.540 & 6/0.51 + ngc 1387 & @xmath0 & * 128 @xmath93 10*&*0.998 @xmath93 0.069 * & 65 - 70 & & 0.000 & 131/0.98 + ngc 1411 & @xmath0 & * 9 @xmath93 0 * & * 0.729 @xmath93 0.002 * & 70 - 74 & 6 & 0.724 & + ngc 1512 & @xmath14 & 46 @xmath93 0 & * 0.605 @xmath93 0.009 * & 140 - 146 & 90 & 0.631 & 56/0.64,*83 * + ngc 1533 & @xmath0 & * 116 @xmath93 5*&*0.942 @xmath93 0.006 * & 55 - 70 & 151 & 0.851 & + ngc 1553 & @xmath0 & * 150 @xmath93 6*&*0.749 @xmath93 0.000 * & 130 - 180 & 150 & 0.633 & /0.93 + ngc 1574 & @xmath0 & * 31 @xmath93 1 * & * 0.960 @xmath93 0.003 * & 95 - 100 & 35 & 0.000 & + ngc 3081 & @xmath14 & 110 @xmath93 10 & * 0.861 @xmath93 0.018 * & 72 - 88 & 158 & 0.776 & * 97/0.837 * + ngc 3100 & @xmath14 & * 148 @xmath93 1*&*0.603 @xmath93 0.012 * & 105 - 115 & 154 & 0.513 & + ngc 3358 & @xmath0 & * 138 @xmath93 0*&*0.581 @xmath93 0.012 * & 64 - 70 & 141 & 0.562 &
125/0.74 + ngc 3706 & @xmath0 & * 76 @xmath93 1 * & * 0.634 @xmath93 0.011 * & 122 - 140 & 78 & 0.603 & + ngc 1079 & 49.3 & 0.268 @xmath93 0.047 & 38.6 & 0.80 & 0.32 + ngc 1317 & 55.1 & 0.065 @xmath93 0.005 & 40.0 & 0.30 & 0.08 + ngc 1317 & 8.7 & 0.100 @xmath93 0.002 & 4.3 & 0.38 & 0.20 + ngc 1326 & 78.3 & 0.157 @xmath93 0.006 & 38.6 & 0.67 & 0.24 + ngc 1326 & 5.8 & 0.040 @xmath93 0.005 & 4.4 & 0.24 & 0.03 + ngc 1350 & 69.6 & 0.172 @xmath93 0.044 & 65.3 & 0.63 & 0.21 + ngc 1387 & 29.0 & 0.067 @xmath93 0.004 & 18.9 & 0.39 & 0.15 + ngc 1533 & 31.9 & 0.105 @xmath93 0.008 & 20.6 & 0.42 & 0.19 + ngc 1553 & & 0.113 @xmath93 0.001 & 31.0 & 0.35 & 0.06 + ngc 1574 & 26.1 & 0.069 @xmath93 0.006 & 13.6 & 0.34 & 0.15 + ngc 3081 & 46.4 & 0.209 @xmath93 0.005 &
34.0 & 0.85 & 0.34 + ngc 3081 & 5.8 & 0.100 @xmath93 0.002 & 4.5 & 0.40 & 0.11 + ngc 3100 & & 0.100 @xmath93 0.006 & 6.0 & 0.21 & + ngc 3358 & 23.2 & 0.100 @xmath93 0.000 & 20.3 & 0.40 & 0.15 + ngc 3358 & 5.8 & 0.170*@xmath93 0.000 & 3.0 & 0.20 & + ngc 718 & 32.2 & 0.121 @xmath93 0.011 & 18.2 & 0.41 & 0.20 + ngc 718 & 13.8 & 0.030 @xmath93 0.000 & 1.1 & 0.15 & + ngc 936 & 59.8 & 0.202 @xmath93 0.036 & 37.0 & 0.58 & 0.30 + ngc 1022 & 27.6 & 0.143 @xmath93 0.009 & 15.9 & 0.41 & 0.14 + ngc 1400 & & 0.034 @xmath93 0.001 & & 0.11 & + ngc 1415 & 71.3 & 0.240 @xmath93 0.046 & 47.2 & 0.80 & 0.25 + ngc 1440 & 27.6 & 0.141 @xmath93 0.004 & 17.7 & 0.50 & 0.23 + ngc 1452 & 46.0 & 0.420 @xmath93 0.044 & 36.1 & 0.90 & 0.58 + ngc 2196 & & 0.067 @xmath93 0.018 & & 0.15 & + ngc 2273 & 41.4 & 0.201 @xmath93 0.013 & 20.0 & 0.57 & 0.25 + ngc 2460 & 9.2 & 0.077 @xmath93 0.028 & 4.4 & 0.14 & 0.04 + ngc 2681 & 27.6 & 0.040 @xmath93 0.000 & 16.1 & 0.23 & 0.06 + ngc 2681 & 4.6 & 0.010 @xmath93 0.000 & 2.3 & 0.10 & 0.03 + ngc 2681 & 64.4 & 0.059 @xmath93 0.001 & 54.5 & 0.30 & 0.07 + ngc 2781 & 62.1 & 0.067 @xmath93 0.000 & 34.5 & 0.46 & 0.10 + ngc 2855 & & 0.075 @xmath93 0.009 & & 0.15 & + ngc 2859 & 69.0 & 0.101 @xmath93 0.003 & 41.6 & 0.57 & 0.22 + ngc 2859 & 6.9 & 0.030 @xmath93 0.000 & 3.5 & 0.15 & 0.03 + ngc 2911 & & 0.081 @xmath93 0.011 & & 0.20 & + ngc 2983 & 41.4 & 0.299 @xmath93 0.027 & 24.6 & 0.77 & 0.45 + ngc 3626 & 39.1 & 0.109 @xmath93 0.003 & 20.5 & 0.35 & 0.08 + ngc 3941 & 29.9 & 0.086
@xmath93 0.002 & 20.0 & 0.34 & 0.14 + ngc 4245 & 50.6 & 0.185 @xmath93 0.007 & 29.2 & 0.54 & 0.20 + ngc 4340 & 78.2 & 0.241 @xmath93 0.016 & 55.9 & 0.55 & 0.27 + ngc 4340 & 4.6 & 0.020 @xmath93 0.000 & 5.8 & 0.05 & + ngc 4596 & 78.2 & 0.276 @xmath93 0.050 & 48.5 & 0.67 & 0.36 + ngc 4608 & 57.5 & 0.251 @xmath93 0.011 & 37.5 & 0.77 & 0.47 + ngc 4643 & 57.5 & 0.299 @xmath93 0.007 & 40.7 & 0.85 & 0.55 + .the decomposition results : @xmath101 is the deviation of @xmath102 from the orientation of the outer disk , whereas @xmath103 is the deviation of @xmath102 from the orientation of the primary bar .
fii for the other three components is the orientation of the component in respect of the outer disk .
[ tbl-4 ] [ cols="<,^,^,^,^,^,^,^,^,^,^,^,^ , < " , ] ll & + galaxy & identifications + eso 208-g21 & bulge / inner disk + ngc 1079 & bulge / disk / bar / extended oval + ngc 1317 & bulge / disk/@xmath104/@xmath105:/@xmath106/@xmath107 + ngc 1326 & bulge / disk/@xmath104/@xmath105/@xmath108/@xmath109 + ngc 1350 & flattened bulge or lens / disk / bar + ngc 1387 & bulge / disk / bar / lens + ngc 1411 & bulge / middle disk / outer disk + ngc 1512 & bulge / disk / bar / oval + ngc 1533 & bulge / disk / bar / lens / inner ring + ngc 1553 & bulge / disk / lens / inner disk + ngc 1574 & bulge / disk / bar / lens + ngc 3081 & bulge / disk/@xmath104/@xmath105/lens + ngc 3100 & middle disk / outer disk + ngc 3358 & bulge / disk/@xmath104/@xmath105/oval + ngc 3706 & bulge / inner disk + ngc 1079 & & 15 & + ngc 1317 & 7 , 13 & 53 & 94 + ngc 1326 & 6 & 28 & 71 + ngc 1350 & & 80 & 170 + ngc 1387 & 6 & & + ngc 1411 & & 37 & + ngc 1512 & 8 & 64 & + ngc 1533 & & 44 & + ngc 1574 & & 17 & + ngc 3081 & 7 & 33 & 57 + ngc 3100 & & & + ngc 3358 & & 22 & + 2d ( final ) & 0.39 @xmath93 0.07 & 2.4 @xmath93 0.2 & 34.7 + 2d ( bulge / disk / bar ) & 0.40 @xmath93 0.07 & 2.3 @xmath93 0.2 & 34.7 + 2d ( bulge / disk ) & 0.61 @xmath93 0.07 & 2.8 @xmath93 0.2 & 34.7 + 1d & 0.55 @xmath93 0.06 & 2.9 @xmath93 0.4 & 25.9 + 2d ( final ) & 0.25 @xmath93 0.03 & 2.1 @xmath93 0.1 & 38.5 + 2d ( bulge / disk / bar ) & 0.30 @xmath93 0.03 & 2.1 @xmath93 0.1 & 38.5 + 2d ( bulge / disk ) & 0.55 @xmath93 0.06 & 2.6 @xmath93 0.1 & 38.5 + 1d & 0.48 @xmath93 0.04 & 2.7 @xmath93 0.4 & 30.2 + @xmath112 & 0.26 @xmath93 0.01 & 0.23 @xmath93 0.02 & 0.18 @xmath93 0.03 & 0.14 @xmath93 0.02 + @xmath113 & 0.41 @xmath93 0.02 & 0.66 @xmath93 0.06 & 0.63 @xmath93 0.06 & 0.49 @xmath93 0.04 + @xmath114 & 0.17 @xmath93 0.01 & 0.32 @xmath93 0.04 & 0.29 @xmath93 0.05 & 0.21 @xmath93 0.03 + @xmath115 & 0.91 @xmath93 0.06 & 1.42 @xmath93 0.11 & 1.72 @xmath93 0.15 & 1.21 @xmath93 0.13 + @xmath116 & 0.57 @xmath93 0.04 & 0.92 @xmath93 0.07 & 1.07 @xmath93 0.11 & 0.80 @xmath93 0.08 + @xmath117 & 57 & 19 & 10 & 10 + | structural analysis has been performed for a sample of 15 southern early - type disk galaxies , mainly s0s , using high resolution @xmath0-band images .
the galaxies are mostly barred and many of them show multiple structures including bars and ovals , typical for s0s .
the new images are of sufficient quality to reveal new detail on the morphology of the galaxies .
for example , we report a hitherto undetected nuclear ring in ngc 1387 , a nuclear bar in ngc 1326 , and in the residual image also a weak primary bar in ngc 1317 . for
the galaxies we ( 1 ) measure the radial profiles of the orientation parameters derived from the elliptical isophotes , ( 2 ) apply fourier methods for calculating tangential forces , and particularly , ( 3 ) apply structural decomposition methods . for galaxies with multiple structures a 2-dimensional method is found to be superior to a 1-dimensional method , but only if in addition to the bulge and the disk , at least one other component is taken into account .
_ we find strong evidence of pseudo - bulges in s0s _ : ten of the galaxies have the shape parameter of the bulge near to @xmath1 = 2 , indicating that the bulges are more disk - like than following the r@xmath2-law .
also , six of the galaxies have either nuclear rings , nuclear bars or nuclear disks . in all non - elliptical galaxies in our sample the @xmath3 0.4 , as typically found in galaxies having pseudo - bulges .
_ in two of the galaxies the @xmath4 flux ratio is as small as in typical sc - type spirals_. this might be the hitherto undiscovered link in the scenario in which spirals are transformed into s0s .
_ also , bars in s0s are found to be shorter and less massive , and have smaller bar torques than bars in s0/a - type galaxies_. |
since the introduction of the k nearest neighbor ( knn ) method by fix and hodges in 1951 @xcite a lot of different variants of it have appeared in order to make it suitable to different scenarios .
the most notable improvements were done in terms of adaptive distance metrics @xcite@xcite@xcite , fast access via space partitioning ( packed r * trees @xcite , kd - trees@xcite , x - trees@xcite , spy - tec @xcite ) , knowledge base ( prototype ) pruning ( @xcite,@xcite ) or classification based on sensitive distributed data ( @xcite,@xcite,@xcite ) .
an overview over the state - of - the - art in nearest neighbor techniques is given in @xcite .
the continuing richness of investigation work into nearest neighbor can be explained with the omnipresence of cbr ( case based reasoning ) type of problems @xcite or just from the practical point of view of its massive parallelizability or simply populartiy .
after all nearest neighbor is conceptually easy to understand - many students get to learn the nearest neighbor as the first classifier .
all of the beforehand mentioned advances evolve around the question of how to optimize the knn s distance measure for retrieving .
the method s apparent laziness might be the reason why a fast preoptimization of @xmath0 has not been paid a lot attention to .
the proper choice of @xmath0 is an important factor for achieving maximum performance of a knn .
however , as we will show , conventional @xmath0 optimization via cross - validation or bootstrapping is slow and promises potential for being sped up .
therefore , we will devote this paper to the concept of fast @xmath0 optimization .
there is work addressing this issue in an alternative way by introducing _
incremental _ knn classifiers based on different types of trees . this class of nearest neighbors attempts to eliminate the influence of k by choosing the right @xmath0 ad hoc .
the rationale behind this method is that for classification tasks the exact majority of a specific class label is not necessarily interesting .
it is only interesting when nearest neighbor is used as a density estimator . in other cases
it is generally enough to feel safe about which class label rules the nearest set .
this class of nearest neighbor starts by polling a minimal amount of nearest neighbors .
then it analyzes the labels and if the retrieved collection of labels is indecisive it will poll more nearest neighbors until it considers the collection decisive enough .
naturally , the method does not scale to small values of @xmath0 , because e.g. a @xmath1 will never be indecisive .
this is a problem because we know from experiments that small @xmath0 are often optimal .
incremental nearest neighbors have their strength in very large databases where typical queries do not need to compute all relative distances . during a lifetime of a large database some distances
might not be computed at all .
the lazy distance computation make incremental nearest neighbor methods ideal candidates for real time tasks operating on large volatile data .
however , in a cross validation setup which needs to compute all distances incremental knns can not play out their strengths . because of the restrictions and intended use we exempted incremental methods from investigation in this paper .
we organize this paper in three parts .
in the first part we will study the options to make the estimation of @xmath0 as fast as possible , in the second part we will experimentally compare the result with the conventional approach and in the last part we will draw a conclusion .
the knn is commonly considered a classifier from the area of supervised learning theory . in this theory
there exists a training function @xmath2 that delivers a model @xmath3 based on a set of options @xmath4 , a matrix of example values @xmath5 and a vector of labels @xmath6 of respective size ( @xmath7 ) . in turn , the model @xmath3 is used in a classification function @xmath8 that is presented with a matrix of new examples @xmath9 and its task is to deliver a vector of new labels @xmath10 ( @xmath11 ) . @xmath2 and @xmath8 must be related but there is no restriction on what the model can be . it can be a set of complex items , a matrix , a vector or just a single value , indeed . from the perspective of a human the purpose of a model
is to make predictions about the future . in order to be able to do this
the human brain requires a simplification of the world . only with the simplification of the world to a reduced number of variables and rules
it can compute future states faster than they occur . in case of knn the model @xmath3
consists of the data and of the smoothing parameter @xmath0 .
this means , that the function @xmath2 is an identity function between model and parameters .
this conflicts with the common notion of a model because the production of models commonly implies reduction . however , the collected data already are a reduction of the world ! from a practical point of view
they can be considered as representative model states of the world and the data in the model is used to predict the class variable of a new vector before it is actually recorded .
this fits perfectly well with the original notion of models .
however , the model is only fixed , when @xmath0 is fixed .
according to the framework , fixing the model ( getting the right value for @xmath0 ) is the job of the function @xmath2 .
many training techniques in machine learning use optimization strategies developed for real numbers and open parameter spaces .
this is not suitable for @xmath0 as it is an integer value and has known left and right limits : an ideal candidate for full search .
most frameworks for pattern recognition offer macro optimization for the remaing model parameters that express themself in the initial training options @xmath4 .
the structural compatibility of the knn with the pattern recognition frameworks macro optimization functions seduces the users very often to macro optimize @xmath0 .
the consequence of this is that knn must compute distances repetitively as it can not assume that specific vectors will simply exchange their role between training and testing in the future . in case of knn
this is exceptionally regrettable .
what does macro optimization mean for the computational complexity ?
here we assume - but without loose of generality - that the dataset @xmath5 is of size @xmath12 ( @xmath12-rows in matrix @xmath5 ) and can be exactly divided into @xmath13 equally sized partitions ready to be rearranged into @xmath13 different train and test setups . since everything is being recomputed the computational complexity for this kind of cross - validation of @xmath0 for a brute knn is @xmath14 .
@xmath15 is the size of the tested range of @xmath0 and since we are considering full search we accept that @xmath15 depends on the size @xmath12 of the dataset and the number of folds @xmath13 .
this means that the @xmath16 .
the scan of nearest sets yields a partial complexity of @xmath17 .
hence , for full search the total complexity is @xmath18 . in order to reduce this high complexity
it is necessary to optimize knn within the train function @xmath2 as it has all necessary information about the relationships among the examples and the labels .
this means that @xmath19 should become @xmath20 where @xmath21 is a vector of partition indices of the kind @xmath22 .
now , the train function @xmath2 can utilize the fact that no new data will arrive during the training and all possible distance requests can be computed in advance .
distances within the same partition need not be computed as they will be never requested .
these fields can be set to infinity ( alternatively they can be filled up with the largest value found in the matrix + 1 ) .
the lower triangle is symmetrical to the upper triangle of the matrix because vectors between two points have the same norm .
the distance matrix @xmath23 has a structure as shown in figure [ fig : initial_matrix ] .
the size of @xmath23 is @xmath24 .
by @xmath25 we mean the component _ distance _ and by @xmath26 we mean the associated _
label_. alone this redesign causes the complexity of the distance computations to get reduced to @xmath27 .
however this benefit is achieved at the expense of higher memory use .
the brute knn has a space complexity of @xmath28 , now the space complexity has risen to @xmath29 .
the next step is to sort the vectors horizontally according to their distance .
although collecting the @xmath0 best solutions would be faster for a single run it means for a range of @xmath0 that you effectively obtain the insertion sort .
since there exist faster sorting algorithms we choose to sort but by using a different algorithm .
the fastest algorithm for doing so is the quick sort .
its average complexity is @xmath30 . in the worst case scenario
the sorting complexity of this method is @xmath31 . in that case
@xmath12 rows will be sorted with @xmath31 .
this means that the worst time complexity so far is @xmath32 and average case is @xmath33 . in order to obtain nearest neighbors for each vector indexed by the row
a counting matrix @xmath34 is initialized with zeros .
@xmath35 is the number of symbols or classes . for each row in @xmath23 and @xmath36 and for the columns @xmath37 in @xmath23 the counters for the specific class label
is increased .
more precisely , for every row @xmath38 and for every tested @xmath0 the counters @xmath39 are updated by @xmath40 .
the complexity of this operation is @xmath41 . for the overall method this adds up to @xmath42 . in parallel the level of correct classification
must be computed because after every modification of @xmath36 the state for the smaller @xmath0 is lost .
therefore a matrix @xmath43 for recording the number of correct classfications is required .
how is this number computed ? at every round @xmath0 of the nearest neighbor candidate computations @xmath44 contains in each of its rows a vector that tells how many labels of specific kind are in the nearest neighbors set .
the classification label is @xmath45 .
the complexity for this operation is @xmath46 .
this simple method is ambiguous by nature , as there can be many labels that are represented by the same amount of vectors in a nearby set .
computer implementations prefer to return the symbol with the smalest coding .
however , it is possible to have a shadow matrix @xmath47 that is the sum of the distances observed for each class label in the set so far .
the rule for computing @xmath48 is the same as for @xmath36 with the difference that instead of adding ones to the matrix you add distances . when symbol frequency is ambiguous ( argmax returns more than one value ) it is possible to use @xmath48 to find which samples are closer overall . because of the specific interest into fast @xmath0 optimization the simple argmax processing is used .
now , every @xmath49 is compared for equality with @xmath50 ( ground truth ) and the binary result is added to @xmath51 .
the @xmath52 will return @xmath13 best @xmath0 .
@xmath53 is obtained by averaging .
considering all parts of the algorithm together the overall complexity is @xmath54
the method for fast @xmath0 computation ( autoknn ) was tested against three other algorithms from the ann library 1.1@xcite : brute , kd - tree and bd - tree knn with default settings .
the autoknn and its competitors performed a complete cross validation run on the ad , diabetes , gene , glass , heart , heartc , horse , ionosphere , iris , mushrooms , soybean , statlog australian , statlog german , statlog heart , statlog sat , statlog segment statlog shuttle , statlog vehicle , thyroid , waveform and wine datasets with 3 , 5 , 10 and 20 folds .
the goal of the experiment was the measurement of the time required to complete the full course of testing different @xmath0 .
the data was separated into stratified partitions which were used in different configurations in order to obtain a training and a testing set .
the autoknn computes the classification results for all @xmath0 while the other algorithms are bound to use a logarithmic search . by logarithmic search
the following schema is meant : @xmath55 .
this schema is practically motivated and rational under the assumption that the influence of additional labels on the result diminishes with higher values of @xmath0 .
practical consideration is primarily test time .
example : while autoknn required 15,35s for a complete scan based on the _ ad _ dataset , on same data exact brute knn needed 353,7s in logarithmic mode and 19595,7s in full mode .
the use of the logarithmic mode makes results with exactly the same values impossible .
however , the differences in resulting @xmath0 and thus in accuracy were absolutely negligible so that the results are directly comparable nonetheless
. we added the experiment times for all databases up to a total for each cross validation size .
the results are shown in the figure and the table under [ fig : result ] .
time measurements were performed on a amd phenom ii 965 with 8 gb of ram with a linux 2.6.35 kernel .
the algorithms are implemented in c / c++ and were compiled with gcc 4.4.5 with o3 option .
only core algorithm operation was measured and all time for additional i / o was ignored . for best comparability , ann library sources were statically included .
the nearest neighbor approach is considered user friendly and is frequently used for data mining , classification and regression tasks .
it is embedded into many automatic environments that make use of knn s flexibility .
although knn has been used , analyzed and advanced for almost six decades a repeating question can not be answered by current literature : what is the fastest way to estimate the right value for @xmath0 and what are the expenses for doing so .
the approach chosen here is to move the @xmath0 esimation away from the meta framework right into the training function @xmath2 .
the advantage of this is that additional information about the data can be made .
this additional information allows to precompute the distances among all vectors without waste and to reuse them numerous times . from this design change which is known to practitioners but not discussed in literature a reduction in time complexity
can be observed from @xmath56 to @xmath57 in average case .
the experiments show , that this has significant impact on the speed of the @xmath0 estimation task .
the comparison between kd - tree knn and the proposed approach proves moreover that having a better time complexity saves practically more time than an efficient distance measure for this task .
the cost of this improvement is a higher space complexity ( now @xmath29 ) . in order to esimtate the practical impact of this complexity exchange we studied the contents of the uci repository @xcite .
the uci should be a reasonable crossover of the problems people face in real life . out of 162 datasets we found that 90% of them have less than 50k examples , 80% of them have less than 10k examples and half of the uci s datasets has less than 1000 examples ( for exact distribution see fig .
[ fig : uci ] ) .
these sizes can be easily handled on higher class commodity computers .
this leads to the conclusion that turning in space complexity for time complexity is a good choice most of the time .
future implementations should offer an integrated @xmath0 searching .
the results also show that the so found values for @xmath0 can be transfered not only to other exact knn but also to approximate knn working on kd - tree and bd - tree models .
this work has been made possible through the funding of the paren ( pattern recognition and engineering @xcite ) project by the bmbf ( federal ministry of education and research , germany ) .
00 fix , e. , hodges , j.l . _ discriminatory analysis , nonparametric discrimination : consistency properties_. technical report 4 , usaf school of aviation medicine , randolph field , texas , 1951 .
carlotta domeniconi , dimitrios gunopulos , jing peng .
_ adaptive metric nearest neighbor classification_. cvpr , vol . 1 ,
pp.1517 ( 2000 ) jing peng , douglas r. heisterkamp , h. k. dai .
_ adaptive kernel metric nearest neighbor classification_. icpr , vol .
3 , pp.30033 ( 2002 ) t. deselaers , r. paredes , e. vidal , h. ney . _
learning weighted distances for relevance feedback in image retrieval_. icpr 2008 , tampa , florida , usa ( 08/12/2008 ) i. kamel , c. faloutsos .
_ on packing r - trees_. proceedings of the 2nd conference on information and knowledge management ( cikm ) , washington dc ( 1993 ) thomas h. cormen , charles e. leiserson , ronald l. rivest .
_ introduction to algorithms_. ch .
10 , mit press and mcgraw - hill ( 2009 ) s. berchtold , d. keim , h .-
_ the x - tree : an index structure for high - dimensional data_. in proceedings of 22th international conference on very large databases ( vldb96 ) , pp 2839 , morgan kaufmann ( 1996 ) dong - ho lee , hyoung - joo kim .
_ an efficient technique for nearest - neighbor query processing on the spy - tec_. ieee transactions on knowledge and data engineering , vol .
15 , no . 6 , pp . 1472 - 1486 , ieee educational activities department , piscataway , nj , usa ( 2003 ) jose salvador - sanchez , filiberto pla , francesco j. ferri . _ prototype selection for the nearest - neighbor rule through proximity graphs_. prl , vol . 18 , no . 6 , pp .
507 - 513 ( 1997 ) u. lipowezky .
_ selection of the optimal prototype subset for 1-nn classification _ prl , vol .
907 - 918 ( 1998 ) keith price bibliography - nearest neighbor literature overview david w. aha . _
the omnipresence of case - based reasoning in science and application_. journal on knowledge - based systems , vol 11 , pp .
261 - 273 ( 1998 ) young , b. , bhatnagar , r. : secure k - nn algorithm for distributed databases .
university of cincinnati ( 2006 ) shaneck , m. , yongdae , k. , kumar , v. : privacy preserving nearest neighbor search . dept . of computer science minneapolis ( 2006 ) kantarcoglu , m. , clifton ,
c. : privately computing a distributed @xmath0-nn classifier .
lncs , vol .
3202 , 279290 ( 2004 ) mount , d. m. : approximate nearest neighbor library 1.1 , department of computer science and institute for advanced computer studies , university of maryland ( 2010 ) uci machine learning repository : pattern recognition and engineering ( paren ) , dfki , bmbf project : ( 2008 - 2010 ) janick v. frasch and aleksander lodwich and faisal shafait and thomas m. breuel _ a bayes - true data generator for evaluation of supervised and unsupervised learning methods_. pattern recognition letters , volume 32:11 , p.1523 - 1531 , 2011 , doi : 10.1016/j.patrec.2011.04.010
the following diagrams are results from knn based on natural and synthetic datasets .
synthetic datasets were obtained using wgks @xcite the standard deviation was estimated based on a 10x cross - validation .
the diagrams are non linear . sections of little change are compressed , hence x - axis are discontinuous . | the k nearest neighbors ( knn ) method has received much attention in the past decades , where some theoretical bounds on its performance were identified and where practical optimizations were proposed for making it work fairly well in high dimensional spaces and on large datasets . from countless experiments of the past
it became widely accepted that the value of @xmath0 has a significant impact on the performance of this method .
however , the efficient optimization of this parameter has not received so much attention in literature .
today , the most common approach is to cross - validate or bootstrap this value for all values in question .
this approach forces distances to be recomputed many times , even if efficient methods are used . hence , estimating the optimal @xmath0 can become expensive even on modern systems .
frequently , this circumstance leads to a sparse manual search of @xmath0 . in this paper
we want to point out that a systematic and thorough estimation of the parameter @xmath0 can be performed efficiently .
the discussed approach relies on large matrices , but we want to argue , that in practice a higher space complexity is often much less of a problem than repetitive distance computations . |
the meson - photon and photon - meson transition form factors contain interesting physics concerning the qcd structure of both photons and mesons .
the pion - photon transition form factor provides a very simple example for the perturbative qcd ( pqcd ) analysis to exclusive processes , and was first analyzed by brodsky and lepage @xcite at large @xmath3 .
it has been shown @xcite that the applicability of pqcd can be extended to lower @xmath3 around a few @xmath5 by taking into the transverse momentum contributions in both hard scattering amplitude and pion wave function . in our recent study @xcite within light - cone quark model , it is shown that the pion - photon transition form factor is identical to the photon - pion transition form factor when taking into account only qcd and qed contributions .
therefore the formalism that applies to the pion - photon transition form factor is also applicable to the photon - pion transition form factor .
taking the minimal quark - antiquark fock states of both the photon and pion as their wave functions , we could calculate the photon - pion transition form factor by using the drell - yan - west assignment .
this framework is applicable at low @xmath3 as a light - cone quark model approach , and it is also physically in accordance with the light - cone pqcd approach at large @xmath3 .
thus we can describe the photon - pion form factors at both low @xmath3 and high @xmath3 within a same framework .
the purpose if this work is to apply this framework @xcite for a systematic description of the photon - meson transition form factors of pseudoscalar mesons @xmath0 , @xmath1 , and @xmath2 , at both @xmath6 and @xmath7 limits , and to make predictions in a wide @xmath3 range .
the photon - meson transition form factor @xmath8 can be realized in @xmath9 or @xmath4 processes .
the @xmath10 transition form factors of @xmath11 , @xmath1 , and @xmath2 at medium to high @xmath3 have been measured at cornell @xcite and at desy @xcite through the @xmath12 process , while the latter process @xmath4 is convenient to provide measurement of the photon - meson transition form factors at low @xmath3 . moreover ,
high precision measurements of the electromagnetic properties of these pseudoscalar mesons via primakoff effect are proposed by primex collaboration at the thomas jefferson national accelerator facility ( jlab ) @xcite , which would give the experimental value of transition form factors @xmath13 of @xmath0 , @xmath1 , and @xmath2 at low @xmath14 , and lead to a clarification on the obvious disagreement between the former primakoff experiment and collider cases in the measurements of @xmath15 and a more precise determination of the @xmath16 mixing angle .
similar measurements can be also performed by hermes collaboration at deutsche elektronen - synchrotron ( desy ) @xcite .
therefore , theoretical predictions at low @xmath17 are necessary and essential for comparison with future experimental measurements .
it is well known that the physical @xmath1 and @xmath2 states dominantly consist of a flavor @xmath18 octet @xmath19 and singlet @xmath20 in the @xmath18 quark model , respectively .
the usual mixing scheme reads : @xmath21 using different sets of experimental data , we recalculate the value of the mixing angle @xmath22 by employing the limiting method developed by cao - signal @xcite .
our results are also compatible with other approaches @xcite for the mixing angle and scheme .
in general , people use the chiral perturbation theory @xcite or some other methods @xcite which deal with current quark masses in order to take the chiral symmetry and chiral anomaly into account , since the chiral symmetry predominates the @xmath23(@xmath24 , @xmath25)@xmath26 vertex at large @xmath17 @xcite , and chiral anomaly determines the @xmath23(@xmath24 , @xmath27 ) transition form factors at @xmath28 ( eqs . ( [ pi_gamma]-[eta_gamma ] ) ) .
in addition , the chiral perturbation theory is also very useful and effective in discussing the @xmath1 and @xmath2 mixing properties @xcite . since we are consistently using the valence quark masses in the light - cone treatment to the form factor calculation
, it is not very applicable to start with current quark mass within the chiral symmetry and investigate the chiral limits in the transition form factor computation .
however , our main purpose of this paper is to employ the new light - cone @xmath29 wave functions @xcite to compute the transition form factors of the light mesons .
moreover , we considered the chiral symmetry when we choose @xmath1 and @xmath2 mixing scheme , and took the chiral limit approximation when we try to determine and fix the parameters
. therefore our results respect the chiral symmetry and its breaking at some extent .
phenomenologically , we could give the predictions of the @xmath1 and @xmath27 mixing angle within the light - cone formalism , as well as the photon - meson transition form factor which is applicable at both low and high energy scales .
this paper is organized as follows . in section 2
we present the formalism for the photon - meson transition form factor using the minimal quark - antiquark fock states of the photon and pion as wave functions . in section 3
, we will introduce the @xmath30 mixing scheme used in our calculation . in section 4 ,
we calculate systematically the photon - meson transition form factors of @xmath31 , @xmath24 , and @xmath32 , and show that the calculated results agree with the available experimental data at medium to large @xmath17 scale .
we also predict the low @xmath3 behaviors of the photon - meson transition form factors of @xmath11 , @xmath1 , and @xmath2 , which are measurable in @xmath4 process via primakoff effect at jlab and desy . in section 5
, we present a brief summary .
we work in the light - cone formalism @xcite , which provides a convenient framework for the relativistic description of hadrons in terms of quark and gluon degrees of freedom , and for the application of perturbative qcd to exclusive processes .
the transition form factor @xmath33 ( @xmath34 , @xmath1 , and @xmath2 ) , in which an on - shell photon is struck by one off - shell photon and decays into a meson , as schematically shown in fig .
[ tranfig ] , is defined by the @xmath35 vertex , @xmath36 in which @xmath37 is the momentum of the off - shell photon , @xmath38 is the squared four momentum transfer of the virtual photon , and @xmath39 is the polarization vector of the on - shell photon .
we choose the light - cone frame @xmath40 [ 0.5 ] .
the arrows indicate the particle moving directions.,title="fig : " ] instead of calculating the diagram directly , we introduce the quark - antiquark wave function of the photon @xcite by calculating the matrix elements of @xmath41which are the numerators of the wave functions corresponding to each constituent spin @xmath42 configuration .
the two boson polarization vectors in light - cone gauge are @xmath43 , where @xmath44 . to satisfy the lorentz condition
@xmath45 the polarizations have the relation @xmath46 with @xmath47 , thus we have @xmath48 & \\ \psi _ { r}^{\uparrow } ( x,\mathbf{k}_{\perp } , \downarrow , \uparrow ) = + \frac { \sqrt{2}(k_{1}+ik_{2})}{x}\varphi _ { \gamma } , & \left [ l^{z}=+1\right ] & \\ \psi _ { r}^{\uparrow } ( x,\mathbf{k}_{\perp } , \uparrow , \uparrow ) = -\frac { \sqrt{2}m}{x(1-x)}\varphi _ { \gamma } , & \left [ l^{z}=0\right ] & \\ \psi _ { r}^{\uparrow } ( x,\mathbf{k}_{\perp } , \downarrow , \downarrow ) = 0 , & & \end{array } \right.\ ] ] in which : @xmath49where @xmath50 is the photon mass and equals to 0 .
each configuration satisfies the spin sum rule:@xmath51 .
therefore , the quark - antiquark fock - state for the photon @xmath52 has the four possible spin combinations : @xmath53 .
\nonumber \\ & & \end{aligned}\ ] ] the quark - antiquark fock - state wave function of the pion is also derived @xcite by using the relativistic field theory treatment of the interaction vertex along with the idea in @xcite . in the light - cone frame of pion ,
@xmath54 we can obtain the four components of the spin wave function by calculating the matrix elements of @xmath55from which we have @xmath56 where @xmath57 is the mass of the quark . after the normalization , we can obtain light - cone representation for the spin structure of the pion , which is the minimal fock - state of the pion light - cone wave function : @xmath58 & \\ \psi _ { \pi l}(x,\mathbf{k}_{\perp } , \downarrow , \uparrow ) = + \frac{m}{\sqrt { 2(m^{2}+\mathbf{k}_{\perp } ^{2})}}\varphi _ { \pi } , & \left [ l^{z}=0\right ] & \\ \psi _ { \pi l}(x,\mathbf{k}_{\perp } , \uparrow , \uparrow ) = + \frac{k_{1}+ik_{2 } } { \sqrt{2(m^{2}+\mathbf{k}_{\perp } ^{2})}}\varphi _ { \pi } , & \left [ l^{z}=-1 \right ] & \\ \psi
_ { \pi l}(x,\mathbf{k}_{\perp } , \downarrow , \downarrow ) = + \frac { k_{1}-ik_{2}}{\sqrt{2(m^{2}+\mathbf{k}_{\perp } ^{2})}}\varphi _ { \pi } , & \left [ l^{z}=+1\right ] & \end{array } \right . \label{lcpionwf}\ ] ] in which we may employ the brodsky - huang - lepage ( bhl ) prescription @xcite , @xmath59 , \label{bhl}\ ] ] for the momentum space wave function , which is a non - relativistic solution of the bethe - salpeter equation in an instantaneous approximation in the rest frame for meson .
each configuration satisfies the spin sum rule : @xmath60 .
hence , the fock expansion of the two particle fock - state for the pion has these four possible spin combinations : @xmath61 .
\nonumber\\ & & \label{wavefunction}\end{aligned}\ ] ] there are two higher helicity @xmath62 components in the expression of the light - cone spin wave function of the pion besides the ordinary helicity @xmath63 components .
such higher helicity components come from the melosh - wigner rotation in the light - cone quark model @xcite , and the same effect plays an important role to understand the proton spin puzzle " in the nucleon case @xcite .
one may also state that these higher helicity components contain contribution from orbital angular moment from a relativistic viewpoint @xcite .
in addition , we would like to add some more remarks on the gaussian - type wavefunction of the bhl prescription that we employ above . as a matter of fact ,
the gaussian wave function is a non - relativistic solution of the bethe - salpeter equation in an instantaneous approximation in the rest frame of the meson as the space wave function .
the bhl wavefunction eq .
( [ bhl ] ) is an extension from a non - relativistic wavefunction into a relativistic form by using the brodsky - huang - lepage ansatz @xcite , and we can consider it as an approximate wavefunction that respects the lorentz invariance in the light - cone formalism .
however , it works phenomenologically well in a lot of calculations ( @xmath64 @xcite ) .
moreover , donnachie , gravelis , and shaw @xcite indicated that the other four possible different space wave functions have similar analytical properties with the bhl wavefunction when the parameter @xmath65 is small ( the @xmath65 is equal to @xmath66 in their paper , the small @xmath65 is corresponding to the @xmath67 and @xmath68 mesons cases ) .
however , they also illustrated that the bhl wavefunction is better than the other four wave functions in the high @xmath65 situation ( for the @xmath69 meson ) .
hence , it gives us the idea that the the bhl wavefunction may be an appropriate choice that we could have right now .
( noticing that the space wave functions for the vector mesons are the same with those for the pseudoscalar mesons , the argument that made by donnachie @xmath70 is also valid for @xmath0 and other pseudoscalar mesons . ) for the physical state of @xmath0 , one should also take into account the color and flavor degrees of freedom into account @xcite @xmath71 , \]]where @xmath72 and @xmath73 are color indices , @xmath74 is the number of colors , and now @xmath75 contains the full spin structure shown above .
so we can get the photon - meson transition form factor of the pion : @xmath76 + ( 1\leftrightarrow 2)\right\ } \\
& = & 4\sqrt{3}(e_{u}^{2}-e_{d}^{2})\int_{0}^{1}\mathrm{d}x\int \frac{\mathrm{d } ^{2}\mathbf{k}_{\perp } } { 16{\pi } ^{3}}\left [ \varphi _ { \pi } ( x,\mathbf{k } _ { \perp } ^{\prime } ) \frac{m}{x\sqrt{m^{2}+\mathbf{k}_{\perp } ^{\prime 2 } } } \right .
\nonumber \\ & & \left .
\times \frac{1}{\frac{m^{2}+\mathbf{k}_{\perp } ^{2}}{x}-\frac{m^{2}+ \mathbf{k}_{\perp } ^{2}}{1-x}}\right ] , \label{fphotonpion}\end{aligned}\]]in which @xmath77 after considering the drell - yan - west assignment @xcite , and @xmath50 @xmath78 is the mass of photon .
in fact , there are two popular mixing schemes for @xmath1 and @xmath2 .
et al _ @xcite suggested the mixing scheme based on the quark flavor basis @xmath80 and @xmath81 , @xmath82 and @xmath83 where @xmath84 is the mixing angle .
the @xmath85 mixing only introduces one mixing angle in the mixing of the decay constants . on the other hand
, people also use the mixing scheme based on the basis of @xmath19 and @xmath20 mixing for @xmath1 and @xmath86 . in the @xmath18 quark model ,
the physical @xmath1 and @xmath2 states dominantly consist of a flavor @xmath18 octet @xmath87 and a singlet @xmath88 @xmath89 , respectively .
the usual mixing scheme reads : @xmath90 in which @xmath22 is the mixing angle . for the calculation of the decay constants of the @xmath19 and @xmath20 mixing ,
feldmann - kroll indicate that two mixing angle scheme could be better from their former investigations@xcite : @xmath91 in addition , one could find that these two schemes could be related by the following equation through the mixing angles finally : @xmath92 from the point view of the chiral symmetry and the @xmath18 symmetry as well as their breaking mechanisms , we find that the @xmath19 and @xmath20 mixing scheme may be more reasonable and physical .
first of all , let us have a brief review on the chiral symmetry and its breaking which have underlying relationship with the @xmath11 , @xmath1 and @xmath2 mesons@xcite . in the chiral symmetry limit , it is well - known that the lagrangian has the @xmath93 symmetry , but the absence of this symmetry in the ground state ( the qcd vacuum ) leads to the chiral symmetry spontaneously breaking into @xmath94 symmetry . because there are 8 spontaneously broken continuous symmetries ( there are 9 when taking into account the chiral anomaly which is associated with the the @xmath95 symmetry breaking ) , there are 8 massless goldstone bosons ( which finally are identified as meson octet ) and 1 massive particle ( which is known as @xmath20 ) according to the goldstone s theorem and chiral anomaly , respectively .
the massless octet includes the meson @xmath96 and @xmath19 .
together with @xmath20 , they mix into massive mesons @xmath11 , @xmath1 and @xmath2 during the explicit @xmath18 symmetry breaking after introducing the quark mass term into the lagrangian . from the above discussion
, we may reach a physical intuitive idea that it is natural and straightforward to use the @xmath20 and @xmath19 mixing scheme as a direct result of the @xmath18 symmetry breaking if we assume that the @xmath96 does not mix with @xmath20 and @xmath19 at all .
from this point of view , the introduction of @xmath20 and @xmath19 is more reasonable than @xmath97 and @xmath98 .
moreover , since it is well - known that pion , kaon , and @xmath19 belong to the same group of octet mesons in the @xmath18 symmetry limit , their parameters should be the same except the quark masses . in this sense
, one may relate the decay constants of @xmath99 and @xmath2 , to pion and kaon in the @xmath100 mixing scheme .
the cleo collaboration@xcite reported their pole fit results as @xmath101 , @xmath102 , and @xmath103 .
these results imply that the nonperturbative properties of @xmath104 and @xmath1 are very similar .
in addition , the absolute value of @xmath22 is small and @xmath105 is the leading order in the @xmath106 mixing scheme of the @xmath1 .
they are consistent with the basic physical intuition that both @xmath104 and @xmath19 are in the @xmath107 octet and are pseudo - massless goldstone particles .
therefore , that is why the authors of @xcite take the parameters of @xmath19 as equal to pion , such as @xmath108 in their paper . from a strict sense ,
if pion , kaon , and @xmath19 are in the same group of octet mesons , the mass of @xmath109 , @xmath110 , and @xmath111 in the bhl wave function should be the same in the calculations of the @xmath104 , @xmath1 and @xmath2 transition form factors .
therefore , we employ the intuitive @xmath112 mixing scheme in the calculations of the @xmath104 , @xmath1 and @xmath86 transition form factors by using the uniform parameters , which shows that the @xmath18 symmetry limit works well in this work . in practice , we utilize the @xmath107 broken form of wave functions for flavor octet @xmath19 and singlet @xmath20 : @xmath113 in which we use gaussian wave function of the bhl prescription : @xmath114 , \\
\phi _ { 8}^{s}(x,\mathbf{k}_{\perp } ) & = & a_{8}\exp \left [ -\frac{m_{s}^{2}+ \mathbf{k}_{\perp } ^{2}}{8\beta _ { 8}^{2}x(1-x)}\right ] , \\ \phi _ { 0}^{q}(x,\mathbf{k}_{\perp } ) & = & a_{0}\exp \left [ -\frac{m_{q}^{2}+ \mathbf{k}_{\perp }
^{2}}{8\beta _ { 0}^{2}x(1-x)}\right ] , \\
\phi _ { 0}^{s}(x,\mathbf{k}_{\perp } ) & = & a_{0}\exp \left [ -\frac{m_{s}^{2}+ \mathbf{k}_{\perp } ^{2}}{8\beta _ { 0}^{2}x(1-x)}\right ] , \end{aligned}\]]and @xmath115 and @xmath81 are the spin parts of the wave functions which are similar to the pion with all possible spin states
. moreover , it is convenient to use the method for the @xmath116 mixing scheme which was developed by cao - signal @xcite in obtaining the mixing angle @xmath22 and the decay constants . in this treatment
, we can get @xmath22 , @xmath117 and @xmath118 directly without involving @xmath119 and @xmath120 . in the @xmath121 mixing scheme
, we have : @xmath122 while for the @xmath0 case , we have : @xmath123 generalizing these equations to @xmath19 and @xmath20 , we have @xmath124 thus we could get : @xmath125 in which we let @xmath126 . along with the same idea by taking the @xmath7 limit
, we could have : @xmath127 in which we have @xmath128 and @xmath129 .
cleo @xcite proposed that the @xmath130 transition form factors could be approximated by : @xmath131 thus we obtain : @xmath132 finally one obtains : @xmath133and gets @xmath134 } , \\
f_{0 } & = & \frac{\sqrt{8}}{4\sqrt{3}\pi ^{2}\left[f_{\gamma \gamma ^{\ast } \rightarrow \eta \prime } ( 0)\sin \theta -f_{\gamma
\gamma ^{\ast } \rightarrow \eta } ( 0)\cos \theta \right ] } , \label{decayconstant}\end{aligned}\ ] ] by using the above results .
there have been many different approaches to discuss the photon - meson transition form factors of light pseudoscalar mesons @xmath31 , @xmath24 , and @xmath32 , such as the light - cone perturbation theory by cao - huang - ma @xcite , the light - front quark model by hwang and choi - ji @xcite , qcd sum rule calculation by radyushkin - ruskov @xcite , and also other approaches _ et al .
we now perform a systematic calculation of these transition form factors in the light - cone framework just presented in section 2 .
the advantage of this new framework is that the predictions should be applicable at both low and high energy scales .
similar to the pion transition form factor and from eq .
( [ eta8 ] ) and eq .
( [ eta0 ] ) , we can get @xmath19 and @xmath20 transition form factors : @xmath136 in which @xmath137 after considering the drell - yan - west assignment , and then we get @xmath138 and @xmath139 in the @xmath19-@xmath20 mixing scheme @xmath140 first of all , we would like to determine the mixing angle @xmath141 and decay constants of @xmath142 and @xmath118 by employing eq .
( [ mixing ] ) to eq .
( [ decayconstant ] ) with two different sets of experimental data which may cast some light on the clarification of the obvious current disagreement between the former primakoff experiments and collider results in the measurements of @xmath143 , and then give more reasonable predictions on the mixing angle @xmath141 . from the particle data group book @xcite
, we get:@xmath144and@xmath145we can get @xmath146 @xmath147 and @xmath148
. however , @xmath149 if we do not include the primakoff production measurement of @xmath150 @xmath151 which obviously disagrees with other collider measurement .
therefore , we obtain @xmath152 @xmath153 and @xmath154 . moreover , we find that the mixing angle @xmath155 is compatible with feld99 which gives the phenomenological value of the mixing angle @xmath156 from eight decay and scattering processes .
the mixing independent ratio @xmath157 can be defined as follow : @xmath158 \label{r1 } \\ & = & \frac{1}{3}\left ( \frac{f_{\pi } ^{2}}{f_{8}^{2}}+8\frac{f_{\pi } ^{2}}{f_{0}^{2}}\right ) .\end{aligned}\ ] ] the current experimental value of @xmath157 which was given in the proposal of the primex collaboration @xcite at jlab is @xmath159 .
we can get @xmath160 and @xmath161 respectively by using the above two sets of the parameters . with the latter set of the fitted value of the mixing angle @xmath22 and decay constants of @xmath142 and @xmath118 as the input parameters
, we can fix the left seven parameters by the following nine constraints . in the formulas of the transition form factor @xmath162 @xmath163 ,
the parameters are the normalization constants @xmath164 , @xmath165 and @xmath166 , the harmonic scale @xmath167 and @xmath168 , and the quark masses @xmath169 and @xmath170 . in order to take a numerical calculation of the transition form factor @xmath171 and compare it with the available experimental data , we need to employ nine constraints to fix those seven parameters above
. thus , we can determine all these seven parameters in the transition form factor uniquely . 1
. the decay widths of @xmath104 , @xmath1 and @xmath2 @xcite : @xmath172 2 .
the @xmath7 limiting behavior of @xmath173 @xcite : @xmath174 3 .
the @xmath7 limiting behavior of @xmath175 @xcite : @xmath176 in which the weak decay constant @xmath177 is defined @xcite from @xmath178 decay .
these constraints are not completely independent , but are necessary since some of them are free from uncertainties , for example , eqs .
( [ constraints5]-[constraints6 ] ) are free from the decay constants @xmath118 and @xmath117 . combined with consideration of other properties of the pion @xcite , we can obtain @xmath179 , @xmath180 , @xmath181 , @xmath182 , @xmath183 , @xmath184 , and @xmath185 . among these 7 parameters , 3 of them ( @xmath109 , @xmath186 and @xmath187 )
are the same in our previous work @xcite and have already been fixed , only the other 4 are new parameters .
these three parameters satisfy eqs .
( [ constraints1 ] ) , ( [ constraints4 ] ) and ( [ constraints7 ] ) very well .
then we fix the 4 new parameters by using the four equations eqs .
( [ constraints2 ] ) , ( [ constraints3 ] ) , ( [ constraints5 ] ) and ( [ constraints6 ] ) . since the parameter fixing scheme is somehow unique , numerical results of these parameters do not have much room to vary , and not surprisingly we find these fixed 7 parameters give very good prediction for eqs .
( [ constraints8])-([constraints9 ] ) .
reversely , we can compute the values of the above nine constraints by using the above seven fixed parameters , and we also provide the fitted values at the end of each equation . therefore , after this simple parameter fixing scheme , we could start to calculate the transition form factor for these mesons .
calculated with the pion wave function in the bhl prescription compared with the experimental data .
the data for the transition form factor are taken from ref .
@xcite and ref .
@xcite . ] compared with the experimental data in the low energy scale .
the data for the transition form factor are taken from ref .
@xcite and ref .
@xcite . ] compared with the experimental data .
the data for the transition form factor are taken from ref .
@xcite , ref .
@xcite and ref .
@xcite . ] compared with the experimental data in the low energy scale .
the data for the transition form factor are taken from ref .
@xcite , ref .
@xcite , ref .
@xcite and ref .
@xcite . ] compared with the experimental data .
the data for the transition form factor are taken from ref .
@xcite , ref .
@xcite , ref .
@xcite and ref . @xcite . ]
the results are in good agreement with the experimental data which we have listed above .
moreover , it is interesting to notice that the masses of the light - flavor quarks ( the up quarks and down quarks ) from the above constrains are just in the correct range ( _ e.g. _ , @xmath188 ) of the constituent quark masses from more general considerations .
naturally , the transition form factor results emerging from this assumption are in quite good agreement with the experimental data . fig .
2 indicates that the theoretical values of the photon - pion @xmath189 transition form factors in the case of low @xmath3 fit the experimental data well .
one may consider this work as a light - cone version of relativistic quark model @xcite , which should be valid in the low - energy scale about @xmath190 .
however , it is also physically in accordance with the light - cone perturbative qcd approach @xcite , which is applicable at the high - energy scale of @xmath191 .
the reason is that the hard - gluon exchange between quark and antiquark of the meson , which should be generally considered at high @xmath3 for exclusive processes , is not necessary to be incorporated in the light - cone perturbative qcd approach for pion - photon transition form factor @xcite . as a result
, there is no wonder that our predictions for the transition form factor at high @xmath3 also agree with the experimental data at high energy scale in fig .
3 . fig . 4 and fig . 5
show that our predictions for the @xmath192 transition form factors agree with the experimental data in the low and high energy scale , respectively .
in addition , the numerical results of @xmath193 transition form factor also give good fit of the experiments both in the low and moderately high energy scale in fig .
6 and fig . 7 .
the prediction that we have made for the low @xmath14 behaviors of the photon - meson transition form factors of @xmath0 , @xmath1 and @xmath2 are measurable in @xmath4 process via primakoff effect at jlab and desy . generally speaking , the medium to high @xmath17 behavior of the transition form factors should include leading - twist order ( so - called pqcd picture ) and nlo corrections @xcite , but we only take the leading order into account in this literature
. however , we find that our results for the leading order of the transition form factors fit the experimental data at small @xmath3 well and are also physically consistent with the light - cone pqcd approach at large @xmath3 .
the light - cone formalism provides a convenient framework for the relativistic description of hadrons in terms of quark and gluon degrees of freedom , and for the application of perturbative qcd to exclusive processes . with the minimal fock - state expansions of the pion and photon wave functions from the light - cone representation of the spin structure of the pseudoscalar meson and photon vertexes , we investigate the photon - meson transition form factors by adopting the drell - yan - west assignment to get the light - cone framework that works at both low @xmath17 and high @xmath17 .
we employ the experimental values of the decay widths of @xmath104 , @xmath1 and @xmath2 , the limiting behavior of @xmath194 @xmath195 , and the limiting behavior of @xmath196 as the nine constrains to fix those seven parameters in the @xmath104 , @xmath197 , and @xmath20 wave functions . with the fixed @xmath104 , @xmath19 , and @xmath20 wave functions
, we find that our numerical predictions for the photon - meson transition form factors of light pseudoscalar mesons @xmath104 , @xmath1 , and @xmath198 agree with the experimental data at both low and high energy scale , in a wide region comparing to previous studies . * acknowledgements * we acknowledge the helpful and encouraging communications with liping gan . this work is partially supported by national natural science foundation of china under grant numbers 10025523 and 90103007
it is also supported by hui - chun chin and tsung - dao lee chinese undergraduate research endowment ( chun - tsung endowment ) at peking university .
brodsky , t. huang , and g.p .
lepage , in _ particles and fields-2 _ , proceedings of the banff summer institute , banff , alberta , 1981 , edited by a.z .
capri and a.n .
kamal ( plenum , new york , 1983 ) , p. 143 . | the photon - meson transition form factors of light pseudoscalar mesons @xmath0 , @xmath1 , and @xmath2 are systematically calculated in a light - cone framework , which is applicable as a light - cone quark model at low @xmath3 and is also physically in accordance with the light - cone pqcd approach at large @xmath3 .
the calculated results agree with the available experimental data at high energy scale .
we also predict the low @xmath3 behaviors of the photon - meson transition form factors of @xmath0 , @xmath1 and @xmath2 , which are measurable in @xmath4 process via primakoff effect at jlab and desy . |
the physics of unstable nuclei is closely related to the equation of state ( eos ) for isospin asymmetric nuclear matter ( ianm ) .
in fact , applications of ianm are broad , ranging from the structure of rare isotopes to the properties of neutron stars . an important quantity that emerges from ianm studies
is the so - called symmetry energy .
however , in spite of many recent and intense efforts , the density dependence of the symmetry energy is not sufficiently constrained by the available data and theoretical predictions show considerable model dependence .
older theoretical studies of ianm can be found in refs .
interactions adjusted to fit properties of finite nuclei , such as those based on the non - relativistic skyrme hartree - fock theory @xcite or the relativistic mean field theory ( see , for instance , ref .
@xcite ) , have been used to extract phenomenological eos . a review of skyrme interactions , particularly popular for nuclear structure applications , can be found in ref .
variational calculations of asymmetric matter were reported in refs .
@xcite , whereas extensive microscopic work with ianm was undertaken by lombardo and collaborators @xcite within the brueckner - hartree - fock ( bhf ) approach .
dirac - brueckner - hartree - fock ( dbhf ) calculations of ianm properties were performed by the oslo group @xcite , the idaho group @xcite , and by fuchs and collaborators @xcite . in this paper , we concentrate on the role of the isovector mesons for the symmetry energy . the latter is defined from an expansion of the energy / nucleon in terms of the isospin asymmetry parameter . in the parabolic approximation , it is simply the difference between the energies per particle in neutron matter and symmetric nuclear matter , see next section . physically ,
it represents the energy price " a nucleus must pay for being isospin asymmetric .
the isovector mesons and their impact on the symmetry energy have been discussed in the literature , particularly , in the context of mean field approaches , both relativistic and non - relativistic ( see , for instance , ref .
@xcite for an extensive review on reaction dynamics with exotic nuclei based on effective interactions based on quantum hadrodynamics ( qhd ) ) .
lately , considerable interest has developed around the symmetry potential , which arises from the difference between neutron and proton single - particle potentials in isospin asymmetric matter . with regard to that issue ,
it is interesting to recall that , in relativistic mean field approaches , introduction of the isovector scalar meson ( the @xmath2 or @xmath3 ) is reported to invert the sign of the splitting between the masses of the neutron and the proton in neutron - rich matter @xcite .
furthermore , in approaches based on qhd , such as those originally proposed by walecka and collaborators @xcite , the dynamical degrees of freedom are essentially included through coupling of the nucleons to the isoscalar scalar @xmath4 and vector @xmath5 mesons .
qhd - i models of nuclear matter do not include the pion , which is perhaps the reason why the contribution of the pion to the symmetry energy may have not been discussed in sufficient depth .
( note , however , that walecka s qhd - ii model does include both @xmath0 and @xmath1 . )
we will explore the role of all isovector channels for the symmetry energy from the point of view of an _ ab initio _ model .
the main point of the _ ab initio _ approach is that mesons are tightly constrained by the free - space data and their parameters are never readjusted in the medium ( this is what we mean by parameter - free " ) .
furhermore , the contributions from the various mesons are fully iterated , thus giving rise to correlation effects .
the corresponding predictions can be dramatically different than those which may be produced in first - order calculations .
this paper is organized as follows : in the next section , we present some facts and phenomenology about the symmetry energy ; then , in section iii , after a brief review of our theoretical approach , we focus on exploring the potential energy contributions of the isovector mesons to the symmetry energy .
our conclusions are summarized in the last section .
asymmetric nuclear matter can be characterized by the neutron density , @xmath6 , and the proton density , @xmath7 , which are related to their respective fermi momenta , @xmath8 or @xmath9 , by @xmath10 with @xmath11 or @xmath12 .
it is more convenient to refer to the total density @xmath13 and the asymmetry ( or neutron excess ) parameter @xmath14 .
clearly , @xmath15=0 corresponds to symmetric matter and @xmath15=1 to neutron matter . in terms of @xmath15 and the average fermi momentum ,
@xmath16 , related to the total density in the usual way , @xmath17 the neutron and proton fermi momenta can be expressed as @xmath18 and @xmath19 respectively .
expanding the energy / particle in ianm with respect to the asymmetry parameter yields @xmath20 where the first term is the energy / particle in symmetric matter and the coefficient of the quadratic term is identified with the symmetry energy , @xmath21 . in the bethe - weizs " acker formula for the nuclear binding energy
, it represents the amount of binding a nucleus has to lose when the numbers of protons and neutrons are unequal . to a very good degree of approximation , one can write @xmath22 the symmetry energy is also closely related to the neutron @xmath23-decay in dense matter , whose threshold depends on the proton fraction .
a typical value for @xmath21 at nuclear matter density ( @xmath24 ) is 30 mev , with theoretical predictions spreading approximately between 26 and 35 mev .
the effect of a term of fourth order in the asymmetry parameter ( @xmath25 ) on the bulk properties of neutron stars is very small , although it may impact the proton fraction at high density .
more generally , non - quadratic terms are usually associated with isovector pairing , which is a surface effect and thus vanishes in infinite matter @xcite .
equation ( [ e ] ) displays a convenient separation between the symmetric and aymmetric parts of the eos , which facilitates the identification of observables that , for instance , may be sensitive mainly to the symmetry energy .
presently , research groups from gsi @xcite , msu @xcite , italy @xcite , france @xcite , and china @xcite are investigating the density dependence of the symmetry energy through heavy - ion collisions .
based upon recent results , these investigations appear to agree reasonably well on the following parametrization of the symmetry energy : @xmath26 where the first term is the kinetic contribution and @xmath27 ( the exponent appearing in the potential energy part ) is found to be between 0.4 and 1.0 . naturally , there are uncertainties associated with all transport models .
recent constraints from msu @xcite were extracted from simulations of @xmath28sn and @xmath29sn collisions with an improved quantum molecular dynamics transport model and are consistent with isospin diffusion data and the ratio of neutron and proton spectra . typically , parametrizations like the one given in eq .
( [ es ] ) are valid at or below the saturation density , @xmath24 .
efforts to constrain the behavior of the symmetry energy at higher densities are presently being pursued through observables such as @xmath30 ratio , @xmath31 ratio , neutron / proton differential transverse flow , or nucleon elliptic flow @xcite .
as stated in the introduction , the starting point of our many - body calculation is a realistic nn interaction which is then applied in the nuclear medium without any additional free parameters .
thus the first question to be confronted concerns the choice of the best " nn interaction . after the development of qcd and the understanding of its symmetries , chiral effective theories @xcite were developed as a way to respect the symmetries of qcd while keeping the degrees of freedom ( nucleons and pions ) typical of low - energy nuclear physics .
however , chiral perturbation theory ( chpt ) has definite limitations as far as the range of allowed momenta is concerned . for the purpose of applications in dense matter ,
where higher and higher momenta become involved with increasing fermi momentum , nn potentials based on chpt are unsuitable .
relativistic meson theory is an appropriate framework to deal with the high momenta encountered in dense matter .
in particular , the one - boson - exchange ( obe ) model has proven very successful in describing nn data in free space and has a good theoretical foundation . among the many available obe potentials , some being part of the high - precision generation " @xcite
, we seek a momentum - space potential developed within a relativistic scattering equation , such as the one obtained through the thompson @xcite three - dimensional reduction of the bethe - salpeter equation @xcite .
furthermore , we require a potential that uses the pseudovector coupling for the interaction of nucleons with pseudoscalar mesons . with these constraints in mind , as well as
the requirement of a good description of the nn data , bonn b @xcite is a reasonable choice .
the mesons included are the pseudoscalar @xmath0 and @xmath32 , the scalar @xmath4 and @xmath2 , and the vector @xmath1 and @xmath5 .
as our many - body framework , we choose the dirac - brueckner - hartree - fock approach .
we will now review the main aspects of our approach and the various approximations we perform through the application of the dbhf procedure .
the main strength of the dbhf approach is its inherent ability to account for important three - body forces ( tbf ) through its density dependence .
these are the tbf originating from virtual excitation of a nucleon - antinucleon pair , known as
z - diagram " .
the characteristic feature of the dbhf method turns out to be closely related to the tbf of the z - diagram type , as we will argue next . in the dbhf approach
, one describes the positive energy solutions of the dirac equation in the medium as @xmath33 where the nucleon effective mass , @xmath34 , is defined as @xmath35 , with @xmath36 an attractive scalar potential .
( this will be derived below . )
it can be shown that both the description of a single - nucleon via eq .
( [ ustar ] ) and the evaluation of the z - diagram generate a repulsive effect on the energy / particle in symmetric nuclear matter which depends on the density approximately as @xmath37 and provides the saturating mechanism missing from conventional brueckner calculations @xcite .
( alternatively , explicit tbf are used along with the bhf method in order to achieve a similar result . ) the approximate equivalence of the effective - mass description of dirac states and the contribution from the z - diagram has a simple intuitive explanation in the observation that eq .
( [ ustar ] ) , like any other solution of the dirac equation , can be written as a superposition of positive and negative energy solutions .
on the other hand , the nucleon " in the middle of the z - diagram is precisely a superposition of positive and negative energy states . in summary , the dbhf method effectively takes into account a particular class of tbf , which are crucial for nuclear matter saturation .
having first summarized the main dbhf philosophy , we now proceed to describe the dbhf calculation of ianm @xcite . in the end
, this will take us back to the crucial point of the dbhf approximation , eq .
( [ ustar ] ) .
we start from the thompson @xcite relativistic three - dimensional reduction of the bethe - salpeter equation @xcite .
the thompson equation is applied to nuclear matter in strict analogy to free - space scattering and reads , in the nuclear matter rest frame , @xmath38 where @xmath39 is the in - medium reaction matrix ( @xmath40=@xmath41 , @xmath42 , or @xmath43 ) , and the asterix signifies that medium effects are applied to those quantities .
thus the nn potential , @xmath44 , is constructed in terms of effective dirac states ( in - medium spinors ) as explained above . in eq .
( [ gij ] ) , @xmath45 , @xmath46 , and @xmath47 are the initial , final , and intermediate relative momenta , and @xmath48 .
the momenta of the two interacting particles in the nuclear matter rest frame have been expressed in terms of their relative momentum and the center - of - mass momentum , @xmath49 , through @xmath50 and @xmath51 the energy of the two - particle system is @xmath52 and @xmath53 is the starting energy .
the single - particle energy @xmath54 includes kinetic energy and potential energy contributions ( see eq .
( [ spe ] ) below ) .
the pauli operator , @xmath55 , prevents scattering to occupied @xmath41 , @xmath42 , or @xmath43 states . to eliminate the angular dependence from the kernel of eq .
( [ gij ] ) , it is customary to replace the exact pauli operator with its angle - average .
detailed expressions for the pauli operator and the average center - of - mass momentum in the case of two different fermi seas can be found in ref.@xcite . with the definitions @xmath56 and
@xmath57 one can rewrite eq .
( [ gij ] ) as @xmath58 & & + \int \frac{d^3k}{(2\pi)^3}v^*_{ij}(\vec q',\vec k ) \frac{q_{ij}(\vec k,\vec p)}{(\epsilon ^*_{ij})_0 -\epsilon
^*_{ij}(\vec p,\vec k ) } g_{ij}(\vec k,\vec q,\vec p,(\epsilon^*_{ij})_0 ) \ , , \label{geq}\end{aligned}\ ] ] which is formally identical to its non - relativistic counterpart .
the goal is to determine self - consistently the nuclear matter single - particle potential which , in ianm , will be different for neutrons and protons . to facilitate the description of the procedure , we will use a schematic notation for the neutron / proton potential .
we write , for neutrons , @xmath59 and for protons @xmath60 where each of the four pieces on the right - hand - side of eqs .
( [ un]-[up ] ) signifies an integral of the appropriate @xmath61-matrix elements ( @xmath41 , @xmath42 , or @xmath43 ) obtained from eq .
( [ geq ] ) .
clearly , the two equations above are coupled through the @xmath43 component and so they must be solved simultaneously .
furthermore , the @xmath61-matrix equation and eqs .
( [ un]-[up ] ) are coupled through the single - particle energy ( which includes the single - particle potential , itself defined in terms of the @xmath61-matrix ) .
so we have a coupled system to be solved self - consistently . before proceeding with the self - consistency
, one needs an _ ansatz _ for the single - particle potential .
the latter is suggested by the most general structure of the nucleon self - energy operator consistent with all symmetry requirements . that is : @xmath62 where @xmath63 and @xmath64 are an attractive scalar field and a repulsive vector field , respectively , with @xmath65 the timelike component of the vector field .
these fields are in general density and momentum dependent .
we take @xmath66 which amounts to assuming that the spacelike component of the vector field is much smaller than both @xmath63 and @xmath67 . furthermore , neglecting the momentum dependence of the scalar and vector fields and inserting eq .
( [ ui2 ] ) in the dirac equation for neutrons / protons propagating in nuclear matter , @xmath68 naturally leads to rewriting the dirac equation in the form @xmath69 with positive energy solutions as in eq .
( [ ustar ] ) , @xmath70 , and @xmath71 the subscript `` @xmath72 '' signifies that these parameters are different for protons and neutrons . as in the symmetric matter case @xcite , evaluating the expectation value of eq .
( [ ui2 ] ) leads to a parametrization of the single particle potential for protons and neutrons ( eqs.([un]-[up ] ) ) in terms of the constants @xmath63 and @xmath67 which is given by @xmath73 also , @xmath74 which , along with eq .
( [ ui3 ] ) , allows the self - consistent determination of the single - particle potential as explained below . the kinetic contribution to the single - particle energy
is @xmath75 and the single - particle energy is @xmath76 the constants @xmath77 and @xmath78 are convenient to work with as they facilitate the connection with the usual non - relativistic framework @xcite .
starting from some initial values of @xmath79 and @xmath80 , the @xmath61-matrix equation is solved and a first approximation for @xmath81 is obtained by integrating the @xmath61-matrix over the appropriate fermi sea , see eq .
( [ ui4 ] ) .
this solution is again parametrized in terms of a new set of constants , determined by fitting the parametrized @xmath82 , eq .
( [ ui3 ] ) , to its values calculated at two momenta , a procedure known as the reference spectrum approximation " .
the iterative procedure is repeated until satisfactory convergence is reached .
finally , the energy per neutron or proton in nuclear matter is calculated from the average values of the kinetic and potential energies as @xmath83 the eos , or energy per nucleon as a function of density , is then written as @xmath84 or @xmath85 clearly , symmetric nuclear matter is obtained as a by - product of the calculation described above by setting @xmath15=0 , whereas @xmath15=1 corresponds to pure neutron matter .
.potential energy contributions ( in mev ) for selected partial waves to the energy of nm and snm , and their difference .
the density is equal to 0.185 @xmath86 . [
cols="^,^,^,^",options="header " , ] in table i we show the contributions of some major partial waves to the potential energy of neutron matter ( nm ) and of symmetric nuclear matter ( snm ) .
the last column displays their difference , to signify the potential energy contribution to the symmetry energy .
the chosen density is 0.185 @xmath86 , corresponding to a fermi momentum of 1.4 @xmath87 in snm and 1.76 @xmath87 in nm ( from eq .
( 3 ) with @xmath15=1 ) .
( summing up all contributions and including the kinetic term yields 33.7 mev , very close to the actual value of our symmetry energy at this density . )
we observe that spin - triplet waves , particularly @xmath88 , give the largest contribution .
it will be interesting to revisit this point in conjunction with the role of the @xmath2 meson .
we note that , although the contribution of the @xmath2 meson to a quantitative nucleon - nucleon ( nn ) interaction is known to be relatively small , this meson is a crucial mechanism to fine - tune the s - waves , that is , @xmath89 _ vs. _ @xmath88 .
hence , its importance for isospin dependent phenomena . in table
ii , we show the contributions to the potential energy of snm from the different mesons
. we show these contributions for potentials a and c as well @xcite .
the three potentials differ mostly in the parameters used for the @xmath90 form factor , which has a large impact on the strength of the tensor force , with bonn a displaying the weakest tensor force and bonn c the strongest ( as demonstrated by the predicted d - state probabilities , which are 4.47 , 5.10 , and 5.53% for a , b , and c , respectively . )
these three potentials span the uncertainty in our knowledge of the short - range tensor force . considering all three models
will then provide information on how the effects being examined ( namely , the role of the isovector mesons on the symmetry energy ) change with changing tensor force , while maintaining consistency with the free - space nn data .
we believe that the latter constraint in crucial for a reliable investigation of many - body effects .
we start by taking @xmath4 and @xmath5 together , since a model with only one of these mesons is entirely meaningless and would produce a completely unrealistic correlated wavefunction , especially so with the @xmath4 alone , due to the absence of any short - range repulsion .
the fourth column is the difference between the values shown in the third and second columns and represents the contribution from the pion alone .
notice that this contribution is attractive , as to be expected , recalling that the pion s tensor potential , @xmath91 , generates a large and attractive second - order term , @xmath92 , when iterated in the bethe - goldstone equation .
consistent with that , this contribution is largest with bonn c , due to its stronger tensor force .
the contribution of the @xmath1 meson is shown in the sixth column as the difference between the values in columns five and three .
it is considerably smaller than the pion s and repulsive , since the tensor force generated by the @xmath1 typically reduces the pion s tensor force at short range . with regard to the @xmath1 , it is useful to recall that the interaction lagrangian which couples vector mesons with nucleons contains both a vector coupling and a tensor coupling , @xmath93 these are related to the electromagnetic properties of the nucleon in the vector dominance model , where the nucleon couples to the photon _ via _ a vector meson . in the framework of the vector dominance model @xcite , a value of 3.7 is obtained for the ratio of the tensor to vector coupling constant , @xmath94 , whereas a stronger value of @xmath95=6.6 was determined from partial - wave analyses @xcite . in other words ,
a larger value of the @xmath1 tensor coupling as compared to its vector coupling is well supported by evidence , a fact that is reflected in meson exchange models where , typically , the ratio @xmath95 is about 6 .
therefore , a lagrangian density with only a vector coupling for the @xmath1 @xcite , _ i. e. _ @xmath96=0 , may miss the most important part of how this meson couples to the nucleon .
back to table ii , the @xmath2 meson is included next , providing a small and positive contribution .
the last column displays the full result , when the pseudoscalar meson @xmath32 is included as well .
table ii is more insightful when examined together with table iii .
the latter shows the same physical quantities as in table ii but for pure neutron matter . here
, the contribution of the pion is much smaller and opposite in sign .
this is due to the absence of the @xmath88 partial wave in nm and , consequently , the absence of a large part of the attractive second - order tensor term mentioned above .
the effect of the @xmath2 meson in nm is of about the same size as the one observed in snm but opposite in sign .
this can be easily understood recalling that the effect of the isovector scalar meson is attractive in @xmath89 and repulsive in @xmath88 , and that the latter is absent from nm . with respect to potential model dependence ,
the size of the effect is largest in model c and weakest in model a. model dependence should be expected , as the parameters of the @xmath2 meson are quite different for the three potentials . before proceeding to discuss the symmetry energy ,
we show , for bonn b , how the various mesons contribute to the energy of symmetric nuclear matter , fig .
1(a ) , and neutron matter , fig .
1(b ) . from fig .
1(a ) , one can see that the effect of the pion is large at all densities . as argued previously ,
this effect comes from the attractive second - order conribution generated by the pion potential , which is clearly quite large already at low density . as density increases ,
the second - order tensor contribution is reduced by the pauli operator ( and dispersion effects ) and thus retains approximately the same size .
we also note the clear impact of the pion on the saturation density of snm , demonstrating the remarkable saturating effect generated by the tensor force , particularly through the @xmath88 partial wave . for neutron matter ,
on the other hand , the contribution of the pion comes mostly from the ( repulsive ) tensor force in some major isospin-1 partial waves .
accordingly , fig .
1(b ) shows that such contribution is opposite in sign and weaker as compared to the one in snm , as already observed when discussing table iii .
also , the effect increases with density , in contrast to the case of snm ; see comments in the previous paragraph . in table iv , we show the difference between the potential energy contributions to nm and snm from the isovector mesons , as an estimate of the effect of each meson on the potential energy part of the symmetry energy .
( the density is the same as in the previous tables . ) clearly , in a microscopic , meson - theoretic approach the impact of the pion on the symmetry energy is the largest .
we find this to be a point of considerable interest , since mean field theories are generally pionless .
this is because the bulk of the attraction - repulsion balance needed for a realistic description of nuclear matter can be technically obtained from @xmath4 and @xmath5 only , an observation that is at the very foundation of walecka models such as qhd - i @xcite .
however , in any fundamental theory of nuclear forces , the pion is the most important ingredient .
chiral symmetry is spontaneously broken in low - energy qcd and the pion emerges as the goldstone boson of this symmetry breaking @xcite .
moreover , nn scattering data can not be described without the pion , which is also absolutely crucial for the two - nucleon bound state , the deuteron .
when moving to nuclear matter ( and regardless the possibility of obtaining realistic values of its bulk properties , including the symmetry energy , with a pionless theory ) , this conceptual problem is not removed .
_ isospin dependence is carried by the isovector mesons : because of their isovector nature , these mesons contribute differently in different partial waves thus giving rise to isospin dependence . _
( this is not the case with isoscalar mesons , which tend to contribute similarly in all partial waves . )
thus , an important aspect of the physics is missing in a discussion of isospin dependence that does not include the pion .
also , conclusions concerning the effect of other mesons ( particularly @xmath1 and @xmath2 ) may be distorted due to the absence of the pion .
this may include , for instance , observations concerning isospin - sensitive quantities such as the neutron - proton mass splitting in neutron - rich matter . as mentioned earlier ,
investigations of @xmath1 and @xmath2 contributions to the potential symmetry energy have been reported , such as the one in refs .
@xcite . in fig
. 6 - 1 of ref .
@xcite , for instance , those contributions are shown to be very large in size ( about -40 mev and 50 mev at saturation density for @xmath2 and @xmath1 , respectively ) .
thus , the interplay between @xmath1 and @xmath2 is described as the equivalent , in the isovector channel , of the @xmath4-@xmath5 interplay in the isoscalar channel @xcite . the dramatic differences between those and
our present observations originate from several sources , which include : the absence of the pion ; the nature of the @xmath1 coupling ; the fact that our meson contributions , when iterated , are reduced by the effect of the pauli projector . as mentioned previously ,
the role of the @xmath2 is important although subtle , and it is found in its different contributions to i=1 and i=0 partial waves , especially the s - waves . in fig .
2 we show the density dependence of the symmetry energy with bonn a , b , and c. the potential model dependence comes almost entirely from differences among predictions of the snm energy . with the three sets of predictions , we mean to estimate the uncertainty to be expected when using different parametrizations for the isovector mesons , while respecting the free - space nn data .
figure 3 displays the momentum dependence of the single - proton and single - nucleon potentials in ianm , as predicted by the three potentials .
differences are small , at most 10% at the lowest momenta .
we recall that the gradient between the potentials shown in fig . 2 , closely related to the isovector optical potential , is the crucial mechanism that separates proton and neutron dynamics in ianm .
in closing this section , we take note of ref .
@xcite , where the effect of the short - range tensor interaction on the symmetry energy is examined using an approximate expression for the second - order tensor contribution @xcite
. it must be noted , though , that the variations performed on the short - range tensor interaction in ref . @xcite are unconstrained and , thus , to some extent arbitrary .
we have examined the effect of the isovector mesons on the difference between the potential energies of pure neutron matter and symmetric matter .
our findings are easily understood in terms of the contributions of each meson to the appropriate component of the nuclear force and the isospin dependence naturally generated by isovector mesons .
we find that the pion gives the largest contribution to this difference .
the contribution of the pion is often ovelooked , possibly because this meson is missing from some mean field models , which are popular among users of equations of state .
it is our opinion that conclusions regarding the interplay of @xmath1 and @xmath2 in phenomenological models must be taken with caution .
we comment on fundamental differences between our approach and the one of mean field models , particularly pionless qhd theories .
first , these differences are of conceptual relevance , since free - space nn scattering and bound state are , essentially , pion physics .
furthermore , they can impact in a considerable way conclusions with regard to isospin dependent systems / phenomena . in order to have a fundamental basis ,
a microscopic theory of the nuclear many - body problem has to start from the bare nn interaction with all its ingredients .
support from the u.s .
department of energy under grant no .
de - fg02 - 03er41270 is acknowledged .
brueckner , s.a .
coon , and j. dabrowski , _ phys .
rev . _ * 168 * , 1184 ( 1968 ) .
siemens , _ nucl .
phys . _ * a141 * , 225 ( 1970 ) .
m. beiner , h. flocard , nguyen van giai , and p. quentin , _ nucl .
phys . _ * a238 * , 29 ( 1975 ) . y. sugahara and h. toki , _ nucl .
phys . _ * a579 * , 557 ( 1994 ) .
j.r . stone and p .-
reinhard , prog .
. phys . * 58 * , 587 ( 2007 ) .
i.e. lagaris and v.r .
pandharipande , _ nucl .
phys . _ * a359 * , 349 ( 1981 ) .
a. akmal , v.r .
pandharipande , and d.g .
ravenhall , _ phys .
c _ * 58 * , 1804 ( 1998 ) .
i. bombaci and u. lombardo , _ phys .
c _ * 44 * , 1892 ( 1991 ) .
w. zuo , i. bombaci , and u. lombardo , _ phys .
c _ * 60 * , 024605 ( 1999 ) .
l. engvik , m. hjorth - jensen , and e. osnes , _ phys .
lett . _ * 73 * , 2650 ( 1994 ) ; l. engvik , e. osnes , m. hjorth - jensen , g. bao , and e. stgaard , _
astrophysical j. _ * 469 * , 794 ( 1996 ) .
d. alonso and f. sammarruca , _ phys .
c _ * 67 * , 054301 ( 2003 ) .
f. sammarruca , int .
e , * 19 * , 1259 ( 2010 ) .
van dalen , c. fuchs , and a. faessler , _ nucl .
phys . _ * a74 * , 227 ( 2004 ) . v. baran , m. colonna , v. greco , and m. di toro , phys
* 410 * , 335 ( 2005 ) .
_ , _ nucl . phys _ * a732 * , 202 ( 2004 ) . j.d .
walecka , ann .
( n.y . ) * 83 * , 491 ( 1974 ) .
b.d . serot and j.d .
walecka , adv .
* 16 * ( 1986 ) .
b.d . serot and j.d .
walecka , int . j. mod .
e * 6 * , 515 ( 1997 ) .
andrew w. steiner , _ phys .
c _ * 74 * , 045808 ( 2006 ) , and references therein .
w. trautmann _
_ , arxiv:0907.2822 [ nucl - ex ] ; c. sfienti _ et al . _
* 102 * , 152701 ( 2009 ) .
lett . _ * 102 * , 12701 ( 2009 ) . v. greco ,
proceedings of the international workshop on nuclear dynamics in heavy - ion reactions and the symmetry energy " , shanghai , china , august 23 - 25 , 2009 , special issue of int . j. phys .
e * 19 * , 1664 ( 2010 ) .
b. borderie and m.f .
rivet , _ prog . part .
phys . _ * 61 * , 551 ( 2008 ) ; arxiv:0812.3524 [ nucl - ex ] .
gao - chan yong , proceedings of the international workshop on nuclear dynamics in heavy - ion reactions and the symmetry energy " , shanghai , china , august 23 - 25 , 2009 , special issue of int .
e * 19 * , 1647 ( 2010 ) .
zhao - qing feng , proceedings of the international workshop on nuclear dynamics in heavy - ion reactions and the symmetry energy " , shanghai , china , august 23 - 25 , 2009 , special issue of int . j. phys .
e * 19 * , 1686 ( 2010 ) .
ko , proceedings of the international workshop on nuclear dynamics in heavy - ion reactions and the symmetry energy " , shanghai , china , august 23 - 25 , 2009 , special issue of int .
e * 19 * , 1763 ( 2010 ) .
s. weinberg , _ physica _ * 96a * ( 1979 ) 327 ; _ phys .
lett . _ * b251 * ( 1990 ) 288 .
r. machleidt and d.r .
entem , phys . rep . * 503 * , 1 ( 2011 ) .
r. machleidt , _ phys .
c _ * 63 * , 024001 ( 2001 ) .
et al . _ , _ phys .
c _ * 49 * , 2950 ( 1994 ) .
thompson , _ phys .
* 1 * , 110 ( 1970 ) .
salpeter and h.a .
bethe , _ phys .
rev . _ * 84 * , 1232 ( 1951 ) .
r. machleidt , _ adv .
nucl . phys . _
* 19 * , 189 ( 1989 ) .
r. brockmann and r. machleidt , _ phys .
b _ * 149 * , 283 ( 1984 ) . m.i .
haftel and f. tabakin , _ nucl .
_ * a158 * , 1 ( 1970 ) .
_ , _ comments nucl . part .
* 17 * , 39 ( 1987 ) .
sakurai , _ currents and mesons _
( university of chicago press , chicago , 1969 ) .
g. h " ohler and e. pietarinen , nucl . phys .
* b95 * , 210 ( 1975 ) .
j. boguta and c.e .
price , nucl .
a505 * , 123 ( 1989 ) .
b. liu , v. greco , v. baran , m. colonna , and m. di toro , phys .
c * 65 * , 045201 ( 2002 ) .
ang li and bao - an li , arxiv:1107.0496 [ nucl - th ] .
brown and r. machleidt , phys .
c * 50 * , 1731 ( 1994 ) . | we examine the potential energy contributions to the symmetry energy ( in the parabolic approximation ) arising from the isovector mesons , @xmath0 , @xmath1 , and @xmath2 .
the significance of a microscopic model which incorporates all important mesons is revealed . in particular , we demonstrate the importance of the pion for a realistic investigation of isospin - sensitive systems . |
the search for any unambiguous signal of quark - gluon plasma ( qgp ) in heavy - ion collisions is motivated by properly investigating the behaviour and properties of hot and dense hg @xcite . here
we formulate a new thermodynamically consistent excluded - volume model where we assign a finite hard - core volume to each baryon but mesons in the theory can easily overlap , fuse and interpenetrate into each other .
earlier , we have used this model successfully in obtaining the conjectured qcd phase boundary and thus determining precisely the location of qcd critical end point @xcite .
we calculate various thermodynamical quantities like number density etc . of hg and compare our model results with that of urasima event generator @xcite .
we use our freeze - out picture @xcite for calculating various hadron ratios and compare our results with the experimental data and various excluded - volume models . in order to make the discussion complete
, we further derive @xmath0 etc . from our model and compare them with other models .
further , we extend our model to deduce the rapidity as well as transverse mass spectra of hadrons and compare them with the experimental data available in order to illustrate the role of flow present in the fluid .
finally , we give summary of this work .
recently , we have proposed a thermodynamically consistent excluded - volume model for a hot , dense hg @xcite . the attractive interaction between baryons and mesons is realized by including the baryon and meson resonances in our model calculation .
the repulsive interaction between baryons is modelled via giving an equal and finite size to each baryon .
mesons can fuse and interpenetrate into each other so , they are treated as pointlike particles . using quantum statistics ,
the grand canonical partition function for baryons can be written as follows : @xmath1}\ ] ] where @xmath2 is the degeneracy factor of ith species of baryons,@xmath3 is the energy of the particle ( @xmath4 ) , @xmath5 is the eigen volume assigned to each baryon of ith species .
apparently , our approach is more simple in comparison to other thermodynamically consistent excluded - volume approach @xcite which often involve transcendental expressions and are difficult to solve .
we determine chemical freeze - out temperature and chemical potential by fitting our results with some experimental data and then use them in determining all other quantities and ratios .
the rapidity distributions of baryons using thermal source can be written as follows @xcite : @xmath6 exp\left(\frac{-m_i\;coshy}{t}\right)\big[m_i^2t+\frac{2m_it^2}{coshy}+\frac{2t^3}{cosh^2y}\big].\ ] ] where @xmath7 is the mass of the ith species .
@xmath8 is the freeze - out volume of the system .
the rapidity spectra of hadrons with the effect of flow can be calculated by using following formula @xcite : @xmath9 where @xmath10 is a free parameter related with the longitudinal flow velocity ( @xmath11 ) @xcite .
our excluded - volume approach involves hard - core repulsion arising between two baryons but mesons do not possess any such repulsion .
we have taken an equal volume @xmath12 for each baryon with a hard - core radius @xmath13 .
we have also taken all baryons , mesons and their resonances having masses upto @xmath14 in our calculation of hg pressure .
we have also imposed the condition of strangeness neutrality by considering @xmath15 where @xmath16 is the strangeness of ith hadron .
dashed line is the result of cleymans - suhonen model @xcite .
symbols are the experimental data @xcite .
rhic data are at mid - rapidity . ] in fig . 1
, we have plotted the variation of number density of hadrons with respect to temperature at fixed net baryon density .
our model results show a very good agreement with the results of sasaki @xcite except at higher @xmath17 .
figure 2 represents the variation of @xmath18 ratio with @xmath19 .
we find that our model calculation gives much better fit to the experimental data at all energies in comparison to cleymans - suhonen model @xcite .
figure 3 depicts the variation of @xmath0 with respect to temperature as obtained in our model for hg @xcite having a baryonic hard - core size @xmath20 fm , and compared the results with those of gorenstein @xmath21 @xcite .
we find that near the expected qcd phase transition temperature ( @xmath22 mev ) , @xmath0 shows a lower value in our hg model than the value in other model . in fig .
4 , we show the rapidity distributions of @xmath23 for central @xmath24 collisions at @xmath25 .
dotted line shows the distribution of @xmath23 due to stationary thermal source .
solid line shows the rapidity distributions of @xmath23 after the incorporation of longitudinal flow in our thermal model and results give a good agreement with the experimental data @xcite . with temperature for @xmath26 in our model and a comparison with the results obtained by gorenstein @xmath21 @xcite .
] at @xmath27 .
dotted line shows the rapidity distribution calculated in our thermal model .
solid line and dashed line show the results obtained after incorporating longitudinal flow in our thermal model .
symbols are the experimental data @xcite . ]
in summary , we find that our model results mostly show very close agreement with those of sasaki , although the two approaches are completely different in nature .
our model calculations for the particle ratios describe the experimental data very well .
transport quantities are also successfully described by our model . the rapidity distributions which essentially are dependent on thermal parameters , also show a systematic behaviour and their interpretations most clearly involve the presence of a collective flow involved in the final description of the fireball .
skt is grateful to council of scientific and industrial research ( csir ) , new delhi for providing a research grant . | we formulate a thermodynamically consistent equation of state ( eos ) , based on excluded - volume approach , for a hot , dense hadron gas ( hg ) .
we calculate various thermodynamical quantities of hg and various hadron ratios and compare our model results with the results of other excluded - volume models and experimental data .
we also calculate various transport coefficients such as @xmath0 etc . and compare them with other hg model results .
furthermore , we test the validity of our model in getting the rapidity spectra of various hadrons and the effect of flow on them is investigated by matching our predictions with the experimental data . * hot and dense hadron gas ( hg ) : a new excluded - volume approach * + * s. k. tiwari * , and * c. p. singh * + department of physics , banaras hindu university , varanasi-05 , india |
recently the technical design report for the superconducting linear collider tesla has been completed @xcite . as a proof that the proposed physics program @xcite can be done with a detector of known technology and reasonable price
the conceptual design of a multipurpose detector has been included @xcite .
the physics goals of tesla require from the detector * very good momentum resolution ( @xmath0 ) e.g. to measure the z recoil mass in the process @xmath1 , * high resolution of the hadronic jet energy ( @xmath2 ) to reconstruct multi - jet events with intermediate resonances such as zhh or @xmath3h , * superb b - tagging to identify multi - b final states like zhh or @xmath3h or to separate @xmath4 and @xmath5 , * optimal hermeticity to reduce backgrounds in missing energy channels , especially to veto two - photon induced events . at tesla
a bunch train consists of around 2800 bunches with a bunch spacing of more than 300 ns .
the relatively long time between bunches makes bunch identification easy and no special fast detectors are needed for this purpose .
the only really relevant background at tesla are @xmath6 pairs created in the collision .
these pairs are concentrated at low angle or low transverse momentum .
the low angle component requires a mask in the forward direction , which can , however , to a large part be used for calorimetry .
the low @xmath7 component at large angles can be kept at low radius by a strong magnetic field .
it limits the radius of the most inner detector layer to around 1.5 cm and causes a significant background in that layer of the vertex detector .
figure [ fig : fuldet ] a ) shows the principle layout of the detector . the tracking system and the calorimeters
are situated inside a 4 t superconducting coil .
sections [ sec : track ] and [ sec : calo ] describe the two main subsystems , tracking and calorimetry . in section [ sec : perf ] the expected performance of the detector is summarised . needed and planned r&d to build the detector presented here is described in other talks of this session .
the layout of the tracking system is shown in figure [ fig : fuldet ] b ) .
it consists of a a precise microvertex detector , a large tpc , silicon tracking in between these two main detectors and a set of forward chambers behind the tpc endplate .
the tpc is in principle similar to existing ones , such as aleph , delphi or star . due to the higher photon background higher
granularity and 100200 pad rows are needed . for this reason
an r&d program has been started to read out the signals with new technologies like gems or micromegas . not to compromise the momentum resolution and the calorimetric energy reconstruction also the field cage and the endplate
will be designed to be as thin as possible . for the microvertex detector ( vtx ) several technologies
are under study , ccds , active pixel sensors and cmos sensors . with the ccds
a point resolution of @xmath8 and a layer thickness of @xmath9 of a radiation length can be reached .
the main problem of ccd detectors for tesla is the relatively long readout time .
reading out the columns in parallel with a frequency of 50mhz results in one complete readout every 100 bunch crossings giving a relatively high but acceptable background in the innermost layer .
for the other technologies an r&d program is in progress to reach a similar performance as presented for the ccds .
the intermediate tracking detector consists of two cylinders of silicon strip detectors in the barrel region down to @xmath10 ( sit ) and three silicon pixel and four silicon strip layers on either side in the forward region ( ftd ) of the detector .
the sit and ftd use only technologies that have already been used successfully in lep so that important r&d is not needed beyond finding ways to thin the detectors .
it improves significantly the momentum resolution in the full acceptance region and enables a precise angle measurement in the forward region before a lot of material is crossed , which is needed to measure the acolinearity of bhabha events in the analysis of the luminosity spectra . to get a reasonably precise momentum measurement below a polar angle of around @xmath11 , where the projected tracklength in the tpc starts to limit the accuracy , at least one precise space point with a resolution around @xmath12 at maximal distance from the interaction point
is needed .
this can for example be done with a forward chamber ( fch ) consisting of 12 straw tube planes with @xmath13 resolution .
the high redundancy helps solving ambiguities in the pattern recognition .
the chamber is also extended over the full tpc endplate to help mapping distortions in the tpc and possibly to serve as a preshower device .
( 18,12)(0,0 ) ( 12,10)(0,-1)1.3 ( 11.3,10.1)lat ( 15,11)(0,-1)1.5 ( 13.5,11.1)tungsten ( 12,2)(3,2)4 ( 10.5,1.4)graphite ( 14.5,1)(3,4)3 ( 13.5,0.4)lcal
to measure jet energies the so called `` energy flow '' concept is proposed . in this scheme
the energy is measured as the sum of the charged particle energies , which contribute about 60% to a typical hadronic jet , the neutral electromagnetic energy measured in the electromagnetic calorimeter , around 30% of a jet , and the energy of neutral hadrons measured in the hadronic calorimeter , representing only the remaining 10% of the energy . in principle
this method allows a very good jet energy resolution , since the tracking resolution is below a few percent , while the resolution of a typical hadron calorimeter is around @xmath14 .
however , in this concept one needs a very good spatial resolution of the calorimeters to separate showers from charged and neutral particles . the setup of the proposed calorimeters can been seen in figure [ fig : fuldet ] . in the barrel and
endcap the electromagnetic ( ecal ) and hadronic ( hcal ) calorimeters , which are inside the coil are followed by the instrumented iron return yoke which is used as a tail catcher and for muon identification . in the very forward region the mask is instrumented for electromagnetic calorimetry . for the ecal two options
are under study , a siw calorimeter and a lead scintillator sandwich with shashlik readout .
the siw option offers several advantages . due to
the high density of tungsten one can pack 25 radiation length into 20 cm thickness . the small moliere radius of tungsten ( @xmath15 ) together with a @xmath16 si - pad size offers a superb lateral resolution and with around 40 layers also a very good longitudinal resolution is possible .
however the detector is fairly expensive ( @xmath17meuro ) . for the shashlik option
a readout granularity of @xmath18 has been studied .
longitudinally a granularity of two is possible using scintillators of different decay times .
the shashlik calorimeter is about a factor six cheaper than for siw , but with worse performance which still has to be quantified .
for the hadronic calorimeter again two options have been studied .
one option is a scintillating tile calorimeter with @xmath19 tiles in the front part and @xmath20 tiles in the back part .
the granularity is limited by the possibility to get the fibres out of the detector . as an alternative
a so called `` digital calorimeter '' has been proposed .
if the granularity of the calorimeter is high enough the energy resolution using only the number of hit cells is better than by using the total amplitude . following this observation as a second possibility
a `` digital calorimeter '' has been proposed , consisting of @xmath21 geiger counters which are read out binary .
the very high granularity is well matched to the siw - ecal allowing an almost perfect separation of showers from charged and neutral particles .
the detailed setup of the mask is shown in figure [ fig : mask ] .
the two parts labelled `` lat '' and `` lcal '' are equipped as calorimeters .
the lat is reasonably clean so that it can be used at least to veto two - photon events .
the lcal receives much background from @xmath6-pairs from beamstrahlung .
probably it is only usable for a fast luminosity measurement for machine tuning .
figure [ fig : mres ] shows the momentum resolution for @xmath22 muons as a function of the polar angle and for particles at @xmath23 as a function of the momentum . over the full tracking region a unique charge identification is possible up to highest momenta and in the central region @xmath24 is reached . with this resolution
the z- and z - recoil mass resolution in @xmath25 is dominated by the z - width and the beam energy spread .
.,title="fig : " ] .,title="fig : " ] with the ccd vertex detector the impact parameter resolution is given by @xmath26 .
for b- and c - tagging the sld package zvtop @xcite has been used .
as detailed in @xcite this allows in z decays at rest to tag b - quarks with 70% efficiency at 95% purity and to tag c - quark with 40% efficiency at 85% purity .
an energy flow resolution of @xmath27 had been reached with the siw - ecal and the digital hcal @xcite . as can be seen from figure [ fig : eflow ] this resolution
is needed for example to measure the higgs self coupling from @xmath28 . a comparable analysis for the shashlik ecal does not exist yet . and background assuming a ) : @xmath29 or b ) : @xmath30 . ]
for the tesla collider a detector has been designed with which the proposed physics from @xmath31 to @xmath32 can be well measured .
the degradation of the physics signals due to the finite resolution is minimal .
the necessary r&d for the proposed detector is explained in other contributions to this session .
the total cost of the detector has been estimated to be between 160meuro and 280meuro depending on the calorimeter choice . | for the superconducting linear collider tesla a multi purpose detector has been designed .
this detector is optimised for the important physics processes expected at a next generation linear collider up to around 1 tev and is designed for the specific environment of a superconducting collider .
snowmass e3029 talk presented at the 2001 snowmass workshop on the future of particle physics july 1 - 20 , 2001 |
heavy baryons provide a testing ground to the standard model ( sm ) .
experiments are accumulating more and more data on heavy baryons @xcite . for the theoretical calculation ,
the task is to apply model - independent methods of the nonperturbative qcd . for the heavy hadrons containing a single heavy quark , the heavy quark effective theory ( hqet ) @xcite
can be applied successfully . after combining the chiral symmetry , the low energy effective theory of hqet , the heavy hadron chiral lagrangian ( hhcl )
@xcite extends the application area further .
recently , the @xmath1 expansion approach @xcite incorporates in more understanding of qcd to the baryon sector .
the heavy baryon decays @xmath0 @xcite are interesting from the point of view of the above mentioned approaches . in the heavy quark limit ,
the most interesting weak transitions @xmath12 and @xmath13 are simplified greatly .
the numbers of the independent form factors are reduced to one and two , respectively in these two transitions .
the decays @xmath14 @xcite however , are predicted to be both @xmath15 and @xmath1 suppressed .
they provide a test to the hqet .
the decays @xmath0 , on the other hand , are neither @xmath15 nor @xmath1 suppressed
. they may be the dominant modes of the decays @xmath16 , although they are phase space suppressed .
we will calculate them in the heavy quark limit .
the calculation involves in the hhcl by assuming the pion being soft .
it also needs the three form factors of @xmath12 and @xmath13 . in the large @xmath11 limit
, there is a light quark spin - flavor symmetry in the baryon sector .
this leads to the fact that only one form factor is independent among the three @xcite .
it is argued that this result is true until to the order of @xmath17 @xcite .
this paper is organized as follows . in sec .
ii , after a brief review of the hhcl , the invariant matrix elements of the decays are calculated in this framework with the help of the @xmath1 expansion .
iii describes necessary kinematics for the phase space integration and presents the numerical results .
the discussion and summary are made in sec .
the dynamics of the decays @xmath0 is shown in fig .
1 , which includes both the strong and the weak interactions .
the strong interaction is described by the hhcl in the case of the pion being soft .
the chiral perturbation theory is the low energy effective theory of qcd for the light pseudoscalar mesons .
these mesons are regarded as the goldstone bosons due to the spontaneous chiral symmetry breaking @xmath18 in qcd .
the eight goldstone boson fields , namely the pion , the kaon and the @xmath19-meson fields , can be represented in the nonlinear exponential form @xcite @xmath20 where @xmath21 and @xmath22 is the pion decay constant , @xmath23 . it is convenient to introduce @xmath24 @xmath25 and @xmath26 transform under the chiral symmetry as @xmath27 where @xmath28 and the special unitary matrix @xmath29 is a complicated nonlinear function of @xmath30 , @xmath31 , and @xmath32 . to the leading order of the derivative expansion , the effective lagrangian for these goldstone bosons is @xmath33 heavy baryons can be included in the chiral lagrangian formalism . the ground state heavy baryons can be classified into two categories .
one is the @xmath34 sextet field @xmath35 which describes the heavy baryons whose light quarks have one unit of angular momentum .
the resulting total angular momentum of these baryons is either 1/2 or 3/2 .
@xmath36 is written as @xcite @xmath37 where @xmath3 is the 4-velocity of the heavy baryon .
@xmath38 is the dirac field and @xmath39 is its spin-3/2 counterpart .
@xmath38 represents the following @xmath40 matrix @xmath41 and @xmath39 has a similar form .
the superscripts denote the charge and the isospin . under the chiral symmetry
, @xmath42 transforms as @xmath43 under the heavy quark spin @xmath44 symmetry , @xmath45 where @xmath46 is the @xmath47 generator and @xmath48 is an infinitesimal parameter .
the other component in the classification is the @xmath34 anti - triplet field @xmath49 corresponding to the baryons whose light quarks have zero angular momentum . the resulting total angular momentum of the baryons is 1/2 . it can be expressed as @xmath50 where the dirac field @xmath51 stands for @xmath52 under the chiral symmetry , it transforms as @xmath53 under the @xmath44 , it transforms just like ( [ su(2 ) ] ) . to the leading order of the derivative expansion , the effective lagrangian of the strong interaction of @xmath42 and @xmath49 with the goldstone bosons , which is invariant under the chiral symmetry and the heavy quark symmetry is @xcite @xmath54 + \delta \bar{s}^{\mu}_{ij } s^{ij}_{\mu } \big\}~ , \label{int}\end{aligned}\ ] ] where @xmath55 s ( @xmath56 ) denote the coupling constants and the covariant derivatives are @xmath57 and @xmath58 the mass difference between the sextet and the anti - triplet baryons is @xmath59 we have neglected the mass differences between the @xmath60 and @xmath61 which is @xmath15 suppressed .
the weak interaction of @xmath62 is described by @xmath63 and the hadronic current in the heavy quark limit can be re - expressed by the matter fields as @xmath64 { \bar s}^\alpha_c(v^\prime)\gamma_\mu\frac{1-\gamma_5}{2 } s^\beta_b(v ) \nonumber\\ & & + \eta(v\cdot v^\prime){\bar t}_c(v^\prime)\gamma_\mu \frac{1-\gamma_5}{2}t_b(v)\big\}~,\end{aligned}\ ] ] where @xmath19 , @xmath65 , and @xmath66 are the isgur - wise functions . and
with the definition @xmath67 , the perturbative qcd correction factor @xmath68 is @xmath69^{-\frac{6}{25 } } \bigg[\frac{\alpha_s(m_c)}{\alpha_s(\mu)}\bigg]^{a_l(v\cdot v^\prime)}~ , \nonumber\\ a_l(w)&=&\frac{8}{27}\big[\frac{w}{\sqrt{w^2 - 1 } } \ln\big(w+\sqrt{w^2 - 1}\big)-1\big]~.\end{aligned}\ ] ] in the large @xmath11 limit , the isgur - wise functions have the following relations @xcite @xmath70 in calculating fig .
1 , the spinor sum relations for the dirac and rarita - schwinger types are used , @xmath71\cdot \frac{1+v\hspace{-2mm}/}{2}~,\end{aligned}\ ] ] where @xmath72 is the dirac spinor and @xmath73 is the rarita - schwinger spinor .
we obtain the squared invariant matrix elements as @xmath74 where @xmath75 and @xmath76 = \frac{16g_f^2}{3f^2}g_3 ^ 2|v_{cb}|^2c_{cb}^2\eta^2 \frac{b}{(p_\pi\cdot v+\delta)^2(p_\pi\cdot v^\prime+\delta)^2}~ , \label{ampb}\ ] ] where @xmath77 in eqs .
( [ ampa]@xmath78[amp2 ] ) , @xmath79 , @xmath80 and @xmath81 are the momentums of the pion , lepton and neutrino , respectively .
the kinematics of the four - body decay has been described in ref .
it is shown in fig .
2 . under the definition of @xmath82
the total decay rate is given by @xmath83 where @xmath84^{\frac{1}{2}}~,\end{aligned}\ ] ] and @xmath85 , @xmath86 are measured in @xmath5 rest frame .
to get the decay rate distributions , the following values are used , @xmath87 the value of @xmath88 is @xmath89 @xcite .
we adopt two forms for @xmath90 .
one is a linear fitting of the result from qcd sum rule @xmath91 @xcite ; and the other is an exponential function from large @xmath11 qcd @xmath92 $ ] @xcite .
the numerical results of @xmath93 and @xmath94 are given in figs .
3 and 4 , respectively .
the decay rate distributions on @xmath67 , pion energy and lepton energy are shown .
the results of the two kinds of isgur - wise function are compared .
the total decay rates are obtained as @xmath95 and hence the branching ratios are @xmath96 for the linear ( exponential ) @xmath90 . compared with the result of @xmath97 @xcite , @xmath98 , and @xmath99 .
they are less dependent on the isgur - wise function because of the partial cancellation of it in the ratios .
we have calculated the decays @xmath0 in the heavy quark , chiral and large @xmath11 limits .
the differential decay spectra are presented .
the branching ratios are obtained as @xmath100 1.21 ( 0.971 ) % , and @xmath101 0.798 ( 0.640 ) % with @xmath90 from qcd sum rule ( large @xmath11 qcd ) .
let us discuss the accuracy of these approximations .
first , consider the finite heavy quark mass correction .
we have taken @xmath102 .
generally the size of the @xmath15 correction can be around @xmath103 .
the detailed consideration indicates that the correction maybe smaller . in the calculation , there are two places where the heavy quark limit is applied .
one is in the weak transition where the quark current was replaced by the baryonic matrix elements . for the @xmath12 transition ,
the @xmath15 correction was calculated to be @xmath104 @xcite . however , there is no estimation for the @xmath15 correction to the @xmath13 transition .
it is experiential to guess that the correction in this case is also at the @xmath105 level .
the other place where the heavy quark limit was used is in the description of the strong interaction , namely in the hhcl .
the difference between the coupling @xmath88 for the bottom quark and that for the charm quark , and the difference between the masses of @xmath60 and @xmath61 , have been neglected .
theoretically these differences are both @xmath15 and @xmath17 suppressed @xcite .
note that practically @xmath106 mev @xcite which is smaller than the pion mass .
hence the correction in the second place is very small @xmath107 . from the above discussion
, we expect that the @xmath15 correction to our results is less than @xmath108 , and is probably around @xmath105 .
second , the validity of the soft pion limit should be addressed .
the calculation is valid in the phase space region where the pion energy is smaller than the chiral symmetry breaking scale which is argued to be at @xmath109 gev @xcite .
therefore for a large part of phase space , the calculation is expected to be reliable .
however , the subleading terms in the derivative expansion in the effective lagrangian have been omitted . and because of the unknown coefficients of the subleading terms , it is still not possible to include them in the calculation .
the correction due to these terms is at the order of @xmath110 or @xmath111 which is less than @xmath112 if the pion is really soft , say @xmath113 gev . for the total decay rate ,
the results we have obtained maybe subject to larger correction . finally , the @xmath1 expansion used in the calculation needs to be explicated .
the calculation needs the information of the three isgur - wise functions of the transitions @xmath12 and @xmath13 . in the large @xmath11 limit ,
the three form factors are related to each other @xcite .
these relations are argued to be valid to the order of @xmath17 @xcite due to the light quark spin - flavor symmetry in the baryon sector in the large @xmath11 limit .
the argument goes like that as in ref .
it is the relation among the form factors , not the knowledge about the form factor itself , that we are considering .
therefore the application of these relations causes @xmath114 uncertainty . up to the above approximations ,
the calculation is in principle model - independent . to get the final numerical results , the model result for the @xmath12 isgur - wise function has to be used
this step brings one more source of uncertainty . according to the model results we adopted ,
the uncertainty is at @xmath108 level .
the decays @xmath0 are indeed experimentally interesting according to our results .
the estimated branching ratios of the decays are not small .
the experiment has been studying the channel @xmath115 @xcite .
a lot of data on @xmath5 can be expected to emerge in the near future from cdf , d0 , hera - b , and lhc - b and so on . with high statistics of the data , the decay channels @xmath0 can be discovered . in the near future
, more reliable theoretical results for the baryonic isgur - wise functions can be obtained , even for that appear in the subleading order of the heavy quark expansion , from lattice hqet or the other model - independent methods . that will give more accurate results of the decay distributions and decay rates .
the uncertainties due to the heavy quark limit , large @xmath11 limit and the model - dependence can be reduced to @xmath105 or smaller . in this case , the precise experimental measurement for @xmath0 will give more information about the validity range of the hhcl .
we would like to thank hai - yang cheng for helpful discussions .
the work was supported in part by kosef through the src program of snu ctp .
jpl and hss also would like to acknowledge the support by the korean research foundation through the 97-sughak program and 1998 - 015-d00054 .
for reviews , see n. isgur and m.b .
wise , in _ b decays _
b. stone , 2nd edition ( world scientific , singapore , 1994 ) , p. 231 ; + m. neubert , phys .
rep . * 245 * , 259 ( 1994 ) ; + i. bigi , m. shifman , and n. uraltsev , ann .
nucl . part .
sci . * 47 * , 591 ( 1997 ) .
wise , phys .
* d45 * , r2188 ( 1992 ) ; g. burdman and j.f .
donoghue , phys .
b280 * , 287 ( 1992 ) ; t .- m .
cheng , c .- y .
cheung , g .-
lin , y .- c .
lin , and h .- l .
yu , phys .
rev . * d46 * , 1148 ( 1992 ) ; erratum : * d55 * , 5851 ( 1997 ) ; for a review , see m.b .
wise , in _
ccast symposium on particle physics at fermi scale _ , ed .
( gordon and breach , 1995 ) . h .- y .
cheng , c .- y .
cheung , w. dimm , g .-
l lin , y.c .
and h .- l .
yu , phys .
* d48 * , 3204 ( 1993 ) ; + c.l.y .
lee , m. lu , and m. b. wise , _ ibid_. * d46 * , 5040 ( 1992 ) ; + j.l .
goity and w. roberts , _ ibid_. * d51 * , 3459 ( 1995 ) ; + g. kramer and w.f .
palmer , phys .
b298 * , 437 ( 1993 ) ; + a. pais and s.b .
treiman , phys . rev . * 168 * , 1858 ( 1968 ) ; + n. cabibbo and a. maksywowicz , _
ibid_. * 137 * , b438 ( 1965 ) ; + g. l. kane , k. stowe , and w.b .
rolnick , nucl . phys . *
b152 * , 390 ( 1979 ) . fig .
3 decay rate distributions for @xmath119 over ( a ) @xmath120 , ( b ) pion energy , and ( c ) lepton energy .
solid lines are the results of linear @xmath90 ( qcd sum rule ) , while the dotted lines are the ones of exponential @xmath90 ( large @xmath11 ) .
all the energy scales are in gev , and the distributions are plotted in units of @xmath121 .
+ .3 cm fig . 4 decay rate distributions for @xmath122 over ( a ) @xmath120 , ( b ) pion energy , and ( c ) lepton energy .
solid lines are the results of linear @xmath90 ( qcd sum rule ) , while the dotted lines are the ones of exponential @xmath90 ( large @xmath11 ) .
all the energy scales are in gev , and the distributions are plotted in units of @xmath121 . | the decay of @xmath0 processes are calculated in the framework of heavy baryon chiral perturbation theory with help of the @xmath1 expansion . the decay rate distributions over @xmath2 where @xmath3 and @xmath4 are the 4-vector velocities of @xmath5 and @xmath6 respectively , and over the pion and lepton energies are given .
the decay branching ratios amount to about @xmath7 for @xmath8 and @xmath9 for @xmath10 by using the baryonic isgur - wise function predicted by qcd sum rule ( large @xmath11 qcd ) . |
the exact cpt invariance implies the identity of the lifetime of a particle and its charge conjugate .
therefore , when the individual partial decay widths of cp conjugate channels are different , due to the cp violation ( cpv ) , the other channels must have equal amount of cp violation , with opposite sign , such that the total width of the particle and antiparticle are equal .
final state interaction must be the responsible to distribute the cp asymmetry among the different conjugate decay channels such as the sum of partial widths provides identical total width for the particle and its anti - particle .
the common belief says that cpt invariance is not a practical constraint to be taken into account when computing cp violation in charmless @xmath0 decays .
it is naively expected that cp violation is distributed among several different coupled final state channels with two , three , four and more hadrons @xcite .
therefore , the cpt constraint should be realized , in principle , by summing up over many hadronic decay channels .
however , since hadronic many - body rescattering effects are far from being understood , it is evident that this phenomenological hypothesis should be better understood before consider its important consequences for the description of the cp violation phenomena . in an opposite direction to
the common believe , we shown in detail in a previous paper @xcite , that the coupling between several final states must be suppressed at least in some charmless three body @xmath0 decays .
a quick observation to the event distribution of these decays shows that they are placed basically at the boundary of the dalitz plot , in a region dominated by light two - body resonances .
this dominance can be directly associated with the non - perturbative regime at the @xmath0 mass energy @xcite .
in addition , by considering the observations made by early experiments @xcite that the low mass @xmath7 and @xmath11 region is dominated by the elastic regime , the number of coupled channel expected for charmless three - body @xmath0 decays should be very restricted .
there is a significant exception to this general picture for the two - body mass distribution : the s - wave @xmath3 re - scattering process between @xmath12 to @xmath8 gev @xcite , that was also discussed in detail in our previous paper .
so , at least for this class of decays , the existence of few possibilities of coupled hadronic channels demands the explicit consideration of the important constraint from cpt invariance in the study of cp violation .
unfortunately little is known , either theoretically or experimentally , about the event distribution for others coupled channels with four or more hadrons in the final state .
however , by looking at the profusion of observed low mass vector - vector charmless @xmath0 decays , it seems that they are also dominated by low mass resonances @xcite .
the inclusive cp asymmetry recently released by the lhcb collaboration for four charged @xmath0 charmless three - body decays , gives a hint about the correlation among the asymmetries in the coupled decay channels @xcite .
actually , the experimental results for strangeness @xmath13 decays @xmath14 and @xmath15 multiplied by the observed number of events @xcite and corrected by the branching fractions ( see ref .
@xcite ) , which represent approximately the efficiency not reported in the paper , show that the number of events involved in the cp asymmetry in one decay channel is approximately the same in the other channel but with opposite sign .
the same happens , within the experimental errors , to the coupled decay channels with strangeness @xmath16 .
taking into account the experimental values @xmath17 and @xmath18 , the number of events observed in each channel , corrected in the same way as before by the branching fractions , show the same correlation as above , namely , a similar number of events involved in the cp asymmetry but with opposite signs . beyond the inclusive cpv
, three - body decays allow to observe this asymmetry in the dalitz phase - space . in principle
, we can expect cp violation in charmless three - body @xmath0 decays coming from three different types of interferences involving weak and strong phases : ( i ) direct cpv due to the interference of the tree and penguin amplitudes in the same intermediate state .
this first kind of cpv is based in the bss model @xcite .
( ii ) cp asymmetry produced through the interference between two different final states with different weak phases coupled by the final state interaction .
this second case is constrained by the cpt invariance and it was already well described in our previous work @xcite .
( iii ) cpv from the interference between two neighbor resonances in the dalitz plot , which share the same - phase space region .
this third cpv kind has been pointed out in previous theoretical .
the present study addresses the three types of dynamics related with cpv in the dalitz plot discussed above and , more specifically , we look for signatures involving each one of these contributions in the charmless decay channels @xmath19 , @xmath6 , @xmath20 and @xmath21 . to perform this study ,
we use the recently published lhcb paper @xcite , in particular , the @xmath22 and @xmath23 mass distribution from the difference between @xmath24 and @xmath25 .
these differences other than give directly the cp violation distribution in the phase - space , minimize substantially problems of acceptance and background not available in the lhcb paper .
our work develops a cp asymmetry formula including resonances and fsi based on the cpt constraint .
the cp asymmetry is derived in lowest order in the strong interaction and decomposed in angular momentum .
the explicit expression for the cp asymmetry is found for the decays channels @xmath5 and @xmath20 .
our discussion is exemplified by considering , within the isobar model , the interference patterns for the cp asymmetry from the @xmath1 and @xmath2 resonances plus a non resonant background including the contribution of the @xmath3 coupled channel amplitude .
we use this formula to fit the recent @xmath19 lhcb data @xcite and obtain as outcome the @xmath6 in qualitative agreement with data for the mass region below @xmath8 gev .
we apply the same procedure to the @xmath26 and @xmath27 decays .
we also take into account the lhcb separation between positive and negative @xmath28 values , that supplies new details which reveal the role of the vector meson resonance in building the cp asymmetry interference pattern .
the work is organized as follows .
a brief review of the cpt constraint for deriving the cp asymmetry with fsi and our notation are provided in sec .
[ seccpt ] .
the cp asymmetry formula in leading order of the strong interaction and angular momentum decomposition is given in sec .
[ sect - cpa ] .
the resonances are introduced in the cp asymmetry expression in sec .
[ sec : rescpt ] .
the formula for the interfering resonant and non - resonant amplitudes in the asymmetry is developed in sec .
[ sec : interf ] .
the analysis of the various cp asymmetry terms is given in sec .
[ sec : analysis ] .
the results from the fitting procedure to the asymmetry in the @xmath29 are shown in sec .
[ sec : cpvfit1 ] , where we found that the cpt violating terms can be disregarded .
we also show results for @xmath30 decay with no new parameters .
an analogous study is performed at sec .
[ sec : cpvfit2 ] for @xmath31 and @xmath32 decays .
the concluding remarks are provided in sec .
[ sec : finalremarks ] , where we also discuss a more general formula to the cp asymmetry , where the fsi terms interfere with the @xmath1 and @xmath2 resonances . in the appendices ,
we show the angle integrated asymmetry as well as the derivation of kinematical factors .
we follow closely refs . @xcite to introduce our notation and the cpt constraint in @xmath0
meson decays .
we start by reminding that the weak and strong hamiltonians conserve cpt , namely , @xmath33 the hadron weak decay amplitude is computed from the matrix element @xmath34 , where the distorted state @xmath35 has the effect of the hadronic strong force due to fsi .
the requirement of cpt invariance on the decay amplitude is @xmath36 where we used that the hadron state @xmath37 transforms under cpt as @xmath38 , where @xmath39 is the charge conjugate state and @xmath40 a phase .
the completeness relation of the strongly interacting states , eigenstates of @xmath41 , and the hermiticity of @xmath42 implies that @xmath43 where @xmath44 defines the matrix elements of the s - matrix .
the sum of the partial widths of the hadron decay channels and the correspondent sum for the charge conjugate are identical , i.e. the mean life of the particle and its conjugate are equal , which is a consequence of eq .
( [ cpt4 ] ) and the hermiticity of @xmath42 ( see ref .
@xcite ) : @xmath45 the cp - violating phase enters linearly at lowest order in the hadron decay amplitude as suggested by the bss mechanism @xcite . in general
, the decay amplitude can be written as @xmath46 where @xmath47 and @xmath48 are complex amplitudes invariant under cp , containing the strongly interacting final - state channel , i.e. , @xmath49 and @xmath50 the only change due to the cp transformation is the sign multiplying the weak phase @xmath51 . the cpt condition depicted by eq .
( [ cpt5 ] ) gives @xmath52 where the subindex labels the final state channels , summed up in the kinematically allowed phase - space . the decay amplitude written in terms of the cpt constraint ( [ cpt4 ] ) , and considering the cp violating amplitudes for the hadron and
its charge conjugate is @xmath53 note that this equation imposes a relation between @xmath47 or @xmath48 with their respective complex conjugates .
the s - matrix is given in terms of the scattering amplitude @xmath54 , namely , @xmath55 which turns eq .
( [ cp11 ] ) into @xmath56 we now consider effect of the fsi in leading order in the decay amplitudes @xmath47 and @xmath48 . for that purpose ,
let us introduce decay amplitudes computed without the effect of the final state interaction , @xmath57 and @xmath58 where @xmath59 and @xmath60 are the mesonic noninteracting charge conjugate states , i.e. , without the distortion of the strong hadronic interaction .
the terms @xmath61 and @xmath62 can be in principle associated with the tree and penguin amplitudes in the bss model @xcite
. the leading order ( lo ) effect due to fsi in the decay amplitude is obtained by substituting @xmath63 and @xmath64 in the right - hand side of eq .
( [ cp11 - 1 ] ) .
considering , the assumption of the cpt invariance of the weak hamiltonian one has also that : @xmath65 which implies in the following relation for the partonic amplitudes @xmath66 and @xmath67 and therefore up to the leading order in @xmath68 , eq .
( [ cp11 - 1 ] ) reduces to @xmath69 which is equivalent to the one provided in refs .
@xcite . in lowest order
the scattering amplitude is restricted only to two - body terms , therefore it is useful to decompose the source decay amplitudes @xmath61 , @xmath70 and the @xmath68 matrix in angular momentum states @xmath71 of each pair in the outgoing channel . the angular momentum decomposition will allow to identify and introduce the meson resonances , like the @xmath72 or @xmath2 in the @xmath7 channel . this leading order decay amplitude ( [ cp24 ] )
decomposed in @xmath71 is written as , @xmath73 where one should note that @xmath74 refers to two - body channels in the final hadronic state including all other dependences on the quantum numbers and in the energy - momentum of the spectator hadrons . by using this expression , the cp asymmetry can be written as @xmath75\right\},\qquad \label{cp26}\end{aligned}\ ] ] where the @xmath76 represents each state coupled by the strong interaction to the decay channel @xmath77 .
we just remind that the second and third terms in the right - hand side of eq .
( [ cp26 ] ) can be associated to the `` compound '' cp asymmetry @xcite .
these two terms cancel each other when summed in all channels @xmath77 and integrated over the phase - space , that leads to the cpt condition expressed by eq .
( [ cp5 ] ) , once the source term in ( [ cp26 ] ) satisfies @xmath78 = 0 , \label{cp22}\end{aligned}\ ] ] which is a consequence of the cpt constraint at the microscopic level , e.g. , as expressed by the tree and penguin amplitudes in the bss model , that should be valid when fsi is turned off in eq .
( [ cpt3 ] ) .
this term was neglected by wolfenstein , which corresponds to the trivial solution of eq .
( [ cp22 ] ) , assuming that the phase difference between the two cp - conserving amplitudes is zero for all decay channels . to be complete and detailing the notation of ref .
@xcite by including the two - particle angular momentum states @xmath71 , we show that the second term in eq .
( [ cp26 ] ) , @xmath79 , \label{cp26d}\end{aligned}\ ] ] also satisfies the cpt condition , namely , this quantity vanishes , which is easily verified by using eqs .
( [ ab01 ] ) as @xmath80 = 0 . \label{cp26d2}\end{aligned}\ ] ] the vanishing of eq .
( [ cp26d2 ] ) is due to the symmetry of @xmath81 , and the fact that @xmath82 , i.e. , the strong interaction does not mix different @xmath83 eigenstates .
therefore , by taking into account eqs .
( [ cp22 ] ) and ( [ cp26d2 ] ) , one has that the cpt constraint @xmath84 \nonumber\\ & + & \sum_\lambda\delta \gamma^{fsi}_\lambda=0 , \label{cpt10}\end{aligned}\ ] ] is fulfilled in leading order of the interaction .
in the case that the channel @xmath77 contains also the formation of a resonance in the partonic process , namely , @xmath85 , the amplitudes @xmath61 and @xmath70 can be separated in the following two parts , @xmath86 , and @xmath87 , where the subindex @xmath88 and @xmath89 mean resonant and non resonant channels .
therefore , the decay amplitude in eq .
( [ cp24 - 1 ] ) is rewritten as @xmath90 \nonumber \\ & + & i\sum_{\lambda^\prime , j}t^j_{\lambda^\prime,\lambda } \left[\sum_r a^j_{0\lambda^\prime r}+a^j_{0\lambda^\prime nr } \right .
\nonumber \\ & + & \left .
b^j_{0\lambda^\prime r}+b^j_{0\lambda^\prime nr}\right ) \right ] .
\label{cp24 - 2}\end{aligned}\ ] ] the resonant source terms @xmath91 and @xmath92 should be interpreted as bare amplitudes , where at the resonance decay vertex , the two - hadron rescattering process is not yet included .
the breit - wigner amplitudes for each term are identified according to @xmath93 and @xmath94 where @xmath71 is the spin of the resonance decaying to two spin zero particles and @xmath95 is the legendre polynomial and @xmath96 is the helicity angle between the equally charge particles in the gottfried - jackson frame
. we will give the representation of this angle for the decay in fig .
[ b - pipipi - theta ] of the next section . after substituting ( [ ress-1 ] ) in ( [ cp24 - 2 ] )
, we get that @xmath97 where the first and second terms in the right - hand side is the isobar model for the decay .
the second term is the source term for the final state channel , and the third one includes the hadronic interaction among the two of the hadrons with angular momentum @xmath71 .
we should clarify that eq .
( [ cp24 - 3 ] ) includes the interaction in the resonance region as the pair of hadrons has a probability to be formed directly from the partonic process . the cp asymmetry from eq .
( [ cp24 - 3 ] ) can be cast in the following form @xmath98^*\ , \left ( \sum _ { r}a_{0\lambda}^{r } f_{r\lambda}^{bw}p_{j}(\cos\theta ) + a^{j}_{0\lambda nr}\right ) \right\ } .
\label{cp27b}\end{aligned}\ ] ] it has to be understood that the subindex @xmath77 also includes different kinematical regions of the three - body channel . since we have introduced the breit - wigner amplitudes in the decay amplitude , the cpt constraint has to be checked in the actual fit , i.e. , if @xmath99 is satisfied when one takes into account the integration over the phase - space besides the sum over all decay channels in the sum of @xmath77 . indeed , in our fitting procedure ,
we will keep only terms that , within our limited model , satisfy the cpt constraint .
we present a simple example to explore the asymmetry formula ( [ cp27b ] ) , considering the resonant , non - resonant source terms and the contribution from the coupling between two strongly interacting channels , namely @xmath7 and @xmath100 .
we use the vector and scalar resonances , the @xmath72 and @xmath2 ones , for instance , interfering with a non resonant amplitude and the term carrying the strong interaction transition amplitude between the coupled channels .
this illustrates exactly the @xmath101 decay case at low invariant @xmath22 mass .
also there are @xmath102 data @xcite previous to the cp asymmetry observation by lhcb collaboration .
first , let us remind that , in a general way , the breit - wigner excitation curve for a resonance @xmath88 reads @xmath103 with @xmath104 being the resonance mass , and @xmath105 denoting the energy dependent relativistic width . for the pion mass , we adopted @xmath106 gev , degenerated for the negative and positive charged particles . here , we consider the resonance decay in the @xmath7 channel .
the real and imaginary parts of @xmath107 are given , respectively , by @xmath108 = \frac{m^2_{r}-s}{(m^2_{r } - s)^2 + m_{r}^2\gamma_{r}(s)^2 } , \end{aligned}\ ] ] and @xmath109 = \frac{m_{r}\gamma_{r}(s)}{(m^2_{r } - s)^2 + m_{r}^2\gamma_{r}(s)^2 } .
\label{real}\end{aligned}\ ] ] the square modulus is @xmath110 the amplitudes for or decays , taking into account the @xmath72 and @xmath2 resonances interfering with a constant non resonant amplitude , can be written as @xmath111e^{\pm i\gamma } , \qquad \label{ampgeral}\end{aligned}\ ] ] where the kinematical factor @xmath112 is included in the amplitude of the @xmath72 vector resonance , to take into account the threshold behavior of the decay amplitude in a @xmath113-wave .
the angle @xmath96 is defined as the angle between the bachelor and the equally charged interacting particle .
see this definition in the decay illustrated in fig .
[ b - pipipi - theta ] .
here , @xmath114 is associated to the spin @xmath12 of the @xmath1 resonance and varies from @xmath115 to @xmath116 along the phase space . being the bachelor particle .
( a ) : @xmath117 ( @xmath118 ) .
( b ) : @xmath119 ( @xmath120 ) . ]
the form factor @xmath121 is included in the non resonant amplitude in order to parametrize the dependence on the square mass of the pair that is brought by the source terms , either for the @xmath7 or @xmath100 systems .
it comes from the partonic decay amplitude that produces the three - meson final state , in which the relative momentum between the pair of mesons is distributed among the quarks in the momentum loop within the microscopic amplitude , e. g. , the tree and penguin diagrams , and probe the internal structure of the mesons involved in the initial and final states . in general
, the non resonant amplitude should depend on the two kinematically independent mandelstam variables , which for simplicity , we choose one of them .
the suggestion is that the microscopic process gives raise to the momentum dependence in a power - law form reflecting the hard momentum structure of the mesons involved in the decay @xcite .
for example , inspecting the tree diagram , one observes that the decay of the @xmath122quark in the rest frame , produces a fast light quark and a pion back to back .
the light quark has to share its momentum with a slow antiquark , which is happens by the gluon exchange and is damped by a momentum power - law form .
a similar reasoning can be applied to the penguin diagram . on the other hand ,
if factorization can be proved , the heavy to light generalized transition form factors @xcite like a heavy meson into @xmath123 , for instance , will also play a role in cpv studies .
an alternative to parametrize the decay amplitude , convenient for monte - carlo simulations , is to write eq .
( [ ampgeral ] ) as @xmath124 where @xmath125 and @xmath126 contain both the fixed weak and strong phases , with the breit - wigner functions introducing additional mass dependent strong phases as sketched above .
the phase @xmath127 comes from the partonic amplitude producing the three - body final state , excluding the strong phase from the rescattering process .
the relation between the parameters is @xmath128 , @xmath129 , and @xmath130 . by comparing the first term in the r.h.s . of eqs .
( [ cp27b ] ) and ( [ ampgeral ] ) , it is possible to make the following identifications , namely , @xmath131 and thus , rewrite eq .
( [ cp27b ] ) for the decay channels @xmath132 or @xmath133 as @xmath134 \nonumber \\ & + & 4(\sin\gamma)\,\mbox{re}\bigg\{\sum_{\lambda^\prime}\bigg [ \left(b_0^\rho f^{\rm bw}_{\rho}k(s)\cos\theta + b_0^f f^{\rm bw}_f + \frac{b^{nr}_{0\lambda}}{1+\frac{s}{\lambda^2_\lambda}}\right)^ * \,t^{j=0}_{\lambda^\prime,\lambda}\ , \frac{a^{nr}_{0\lambda^\prime}}{1+\frac{s}{\lambda^2_{\lambda^\prime}}}\ , \nonumber\\ % & -&\left(\frac{b^{nr}_{0\lambda^\prime}}{1+\frac{s}{\lambda^2_{\lambda^\prime}}}\,\,t^{j=0 } _ { \lambda^\prime,\lambda}\right)^*\,\ , \left(a_0^\rho f^{\rm bw}_{\rho}k(s)\cos\theta + a_0^f f^{\rm bw}_f + \frac{a^{nr}_{0\lambda}}{1+\frac{s}{\lambda^2_\lambda}}\right)\bigg]\bigg\ } , \label{cp27d}\end{aligned}\ ] ] where we used the identity @xmath135 and kept only one term in the scattering amplitude for @xmath136 in eq .
( [ cp27b ] ) .
the subindex @xmath77 are associated with a position in the phase - space of the @xmath0 decay and @xmath76 is equal to @xmath137 or @xmath138 , in the case of @xmath139 and @xmath140 , respectively . the second term in the r.h.s . of eq .
( [ cp27d ] ) is the compound contribution to the cp asymmetry that includes the effect of the final state interaction in the non resonant channel . in this example , only the @xmath141-wave amplitude @xmath142 channel was included and we have considered the interference of this term with the resonant ones .
we use here the functional form of this term obtained in ref .
@xcite .
the compound contribution to cpv , namely , the one carrying @xmath143 in eq .
( [ cp27d ] ) , has as input in our calculations the non - diagonal scattering amplitude depicted in the full s - matrix in the isoscalar and angular momentum zero @xmath144 coupled channels , which is written as @xmath145 with the inelasticity parameter @xmath146 and the @xmath7 phase - shift @xmath147 , given by @xcite , @xmath148 with @xmath149 and @xmath150 - 1}{\cot^2[\delta_{\pi\pi}(s ) ] + 1}\bigg\},\end{aligned}\ ] ] with @xmath151 respectively . in these expressions , we use @xmath152 gev , @xmath153 gev , @xmath154 gev , @xmath155 gev , @xmath156 , @xmath157 , and @xmath158 , according to the parametrization given in ref .
the off - diagonal term in the s - matrix gives the transition amplitude between the @xmath7 and @xmath100 channels in the isoscalar channel , namely , @xmath159 where we have made the approximation @xmath160 between @xmath12 and @xmath8 gev .
the expression for the cp asymmetry in eq .
( [ cp27d ] ) , should be expanded in a form where the unknown parameters are exposed in a simple manner to proceed with the fitting to the experimental data . by using the relations @xmath161 , @xmath162 , and @xmath163 , together with eqs .
( [ real])-([modulus ] ) , one can finally write eq .
( [ cp27d ] ) as @xmath164 + \mathcal{b}^\prime\sin[2\delta_{\pi\pi}(s)]\right\}\sqrt{1-\eta^2(s)}}{\left(1+\frac{s } { \lambda^2_{\lambda}}\right)\left(1+\frac{s } { \lambda^2_{\lambda^\prime}}\right ) } % \nonumber \\ % + \mathcal{c}|f^{\rm bw}_\rho(s)|^2k^2(s)\cos^2\theta \nonumber \\ & + & |f^{\rm bw}_\rho(s)|^2k(s)\cos\theta \left\{\frac{\mathcal{d}(m^2_\rho - s)}{1+\frac{s } { \lambda^2_{\lambda } } } + \frac{\mathcal{d}^\prime\sqrt{1-\eta^2(s)}\left\{m_\rho\gamma_\rho(s)\cos[2\delta_{\pi\pi } ( s)]-(m^2_\rho - s)\sin[2\delta_{\pi\pi}(s)]\right\}}{1+\frac{s}{\lambda^2_{\lambda^\prime } } } \right .
% & + & \left .
\frac{\mathcal{e}m_\rho\gamma_\rho(s)}{1+\frac{s } { \lambda^2_{\lambda } } } + \frac{\mathcal{e}^\prime\sqrt{1-\eta^2(s)}\left\{(m^2_\rho - s)\cos[2\delta_{\pi\pi } ( s)]+m_\rho\gamma_\rho(s)\sin[2\delta_{\pi\pi}(s)]\right\}}{1+\frac{s } { \lambda^2_{\lambda^\prime}}}\right\ } \nonumber \\ % & + & |f^{\rm bw}_{\rho}(s)|^2|f^{\rm bw}_f(s)|^2k(s)\cos\theta\times \nonumber\\ & \times&\left\{\mathcal{f}[(m_\rho^2 -s)(m_f^2-s ) + m_\rho\gamma_\rho(s)m_f\gamma_f(s ) ] + \mathcal{g}[(m_\rho^2 -s)m_f\gamma_f(s ) - m_\rho\gamma_\rho(s)(m_f^2-s)]\right\ } \nonumber \\ % & + & |f^{\rm bw}_f(s)|^2 \left\{\frac{\mathcal{h}(m^2_f - s)}{1+\frac{s}{\lambda^2_{\lambda } } } + \frac{\mathcal{h}^\prime\sqrt{1-\eta^2(s)}\left\{m_f\gamma_f(s)\cos[2\delta_{\pi\pi } ( s)]-(m^2_f - s)\sin[2\delta_{\pi\pi}(s)]\right\}}{1+\frac{s}{\lambda^2_{\lambda^\prime } } } \right .
% & + & \left .
\frac{\mathcal{p}m_f\gamma_f(s)}{1+\frac{s } { \lambda^2_{\lambda } } } + \frac{\mathcal{p}^\prime\sqrt{1-\eta^2(s)}\left\{(m^2_f - s)\cos[2\delta_{\pi\pi } ( s)]+m_f\gamma_f(s)\sin[2\delta_{\pi\pi}(s)]\right\}}{1+\frac{s}{\lambda^2 _ { \lambda^\prime}}}\right\ } + \mathcal{q}|f^{\rm bw}_f(s)|^2 , \label{dgamacos } \end{aligned}\ ] ] where @xmath77 is associated to @xmath165 or @xmath166 , and @xmath76 to @xmath167 or @xmath168 .
the masses of the resonances are @xmath169 gev , and @xmath170 gev @xcite .
the widths used in eq .
( [ ress ] ) for @xmath171 are @xmath172 gev and @xmath173 gev
. the parameters of the model can be written in terms of those in eqs .
( [ ampgeral ] ) or ( [ ampgeral - mcarlo ] ) as follows , @xmath174 = ( a^{nr}_{+\lambda})^2 - ( a^{nr}_{-\lambda})^2,\end{aligned}\ ] ] @xmath175,\end{aligned}\ ] ] @xmath176,\end{aligned}\ ] ] @xmath177 = ( a^{\rho}_+)^2 - ( a^{\rho}_-)^2,\end{aligned}\ ] ] @xmath178 \nonumber \\ & = & % = 2[a^{\rho}_+a^{nr}_{+\lambda } \cos(\delta^{\rho}_+ - \delta^{nr}_{+\lambda } ) - a^{\rho}_-a^{nr}_{-\lambda}\cos(\delta^{\rho}_- - \delta^{nr}_{-\lambda})],\qquad\end{aligned}\ ] ] @xmath179 \nonumber \\ & = & % = 2[a^{\rho}_+a^{nr}_{+\lambda^\prime } \cos(\delta^{\rho}_+ - \delta^{nr}_{+\lambda^\prime } ) % \nonumber \\ & - & -a^{\rho}_-a^{nr}_{-\lambda^\prime}\cos(\delta^{\rho}_- - \delta^{nr}_{-\lambda^\prime})],\quad\,\,\,\end{aligned}\ ] ] @xmath180 \nonumber \\ & = & % = -2[a^{\rho}_+a^{nr}_{+\lambda } \sin(\delta^{\rho}_{+ } - \delta^{nr}_{+\lambda } ) - a^{\rho}_-a^{nr}_{-\lambda}\sin(\delta^{\rho}_- - \delta^{nr}_{-\lambda})],\qquad\end{aligned}\ ] ] @xmath181 \nonumber \\ & = & % = 2[a^{\rho}_+a^{nr}_{+\lambda^\prime } \sin(\delta^{\rho}_+ - \delta^{nr}_{+\lambda^\prime } ) + a^{\rho}_-a^{nr}_{-\lambda^\prime}\sin(\delta^{\rho}_- - \delta^{nr}_{-\lambda^\prime})],\quad\,\,\,\end{aligned}\ ] ] @xmath182 \nonumber\\ & = & % = 2[a^{\rho}_+a^{f}_+ \cos(\delta^{\rho}_+ - \delta^{f}_+ ) - a^{\rho}_-a^{f}_- \cos(\delta^{\rho}_- - \delta^{f}_-)],\qquad\end{aligned}\ ] ] @xmath183 \nonumber \\ & = & % = 2[a^{\rho}_+a^{f}_+ \sin(\delta^{\rho}_+ - \delta^{f}_+ ) - a^{\rho}_-a^{f}_- \sin(\delta^{\rho}_- - \delta^{f}_-)],\qquad\end{aligned}\ ] ] @xmath184 \nonumber \\ & = & % = 2[a^{f}_+a^{nr}_{+\lambda } \cos(\delta^{f}_+ - \delta^{nr}_{+\lambda } ) -a^{f}_-a^{nr}_{-\lambda}\cos(\delta^{f}_- - \delta^{nr}_{-\lambda})],\qquad\end{aligned}\ ] ] @xmath185 \nonumber \\ & = & % = 2[a^{f}_+a^{nr}_{+\lambda^\prime } \cos(\delta^{f}_+ - \delta^{nr}_{+\lambda^\prime } ) - a^{f}_-a^{nr}_{-\lambda^\prime}\cos(\delta^{f}_- - \delta^{nr}_{-\lambda^\prime})],\quad\,\,\,\end{aligned}\ ] ] @xmath186 \nonumber \\ & = & % = -2[a^{f}_+a^{nr}_{+\lambda } \sin(\delta^{f}_{+ } - \delta^{nr}_{+\lambda } ) -a^{f}_-a^{nr}_{-\lambda}\sin(\delta^{f}_- - \delta^{nr}_{-\lambda})],\quad\,\end{aligned}\ ] ] @xmath187 \nonumber \\ & = & % = 2[a^{f}_+a^{nr}_{+\lambda^\prime } \sin(\delta^{f}_+ - \delta^{nr}_{+\lambda^\prime } ) + a^{f}_-a^{nr}_{-\lambda^\prime}\sin(\delta^{f}_- - \delta^{nr}_{-\lambda^\prime})],\quad\,\,\,\end{aligned}\ ] ] and @xmath188 = ( a^{f}_+)^2 - ( a^{f}_-)^2.\end{aligned}\ ] ] in the previous work of ref .
@xcite , we used a different form for the @xmath189 amplitude , writing it as @xmath190 , where @xmath191 and @xmath192 .
this form is analogous of that used in eq .
( [ dgamacos ] ) , in terms of @xmath193 and @xmath194 .
the phase @xmath195 was chosen to be zero that is the same as to take @xmath196 , assumed hereafter . as a remark , we call the reader attention that for the compound contribution , we have discarded three - body rescattering effects at the two - loop level since they are small compared to the first two - body collision contribution , as suggested by the three - body model calculations for the @xmath197 decay , see refs . @xcite .
in order to be able to compare our model directly with the recent experimental data regarding the and decays presented in ref .
@xcite , we had to eliminate one variable in eq .
( [ dgamacos ] ) , namely , the square mass of the pair presenting the bachelor particle .
we perform this calculation in appendix [ app - int ] . here in this section
, we show the behavior of some terms of the integrated eq .
( [ dgamacos ] ) . as it was mentioned before , the term of the integrated eq .
( [ dgamacos ] ) containing the @xmath198 parameter , is equivalent to the direct cpv induced by the interference between the tree and penguin amplitudes with one meson resonance in the final state . in this case , the contribution from cpv comes only from the @xmath72 meson . in fig .
[ dcpv ] , we show the signature of the cpv from this term with the usual breit - wigner square modulus and with the same sign for both cases , namely , @xmath117 and @xmath119 .
this term locally violates the cpt constraint and will be set to zero in our fittings , by assuming @xmath199 .
parameter of the integrated eq .
( [ dgamacos ] ) , as a function of @xmath200 for ( a ) @xmath117 and ( b ) @xmath119 . in these plots we set @xmath201 . ] the dalitz cpv induced by the interference term of the integrated eq .
( [ dgamacos ] ) , namely , that containing the @xmath202 parameter , is directly related with the real part of the breit - wigner function .
this function presents a clear signature in the mass spectrum , namely , a zero and a change in the sign of the cp asymmetry at the central value of the @xmath72 mass .
another peculiarity associated to vector mesons resonances is one more sign change when the @xmath28 cross the zero around the middle of the dalitz plot .
these two facts can be observed in fig .
[ dcpv - real ] .
notice that this term sums to zero and thus locally satisfies the cpt constraint .
parameter of the integrated eq .
( [ dgamacos ] ) , as a function of @xmath200 for ( a ) @xmath117 and ( b ) @xmath119 .
here we have used @xmath203 gev and @xmath204 . ] the term containing the @xmath205 parameter in the integrated eq .
( [ dgamacos ] ) , associated to the dalitz cpv induced by the interference between the @xmath72 and the non resonant amplitude , is directly related with the imaginary part of the breit - wigner function of @xmath72 .
the shape is similar to the those presented in fig .
[ dcpv ] , with the clear difference that the proportionality with @xmath28 changes the sign of the cp asymmetry when @xmath28 pass through zero in the middle of the dalitz plot .
[ dcpv - imag ] shows these features .
notice that this term sums to zero and thus locally satisfies the cpt constraint .
parameter of the integrated eq .
( [ dgamacos ] ) , as a function of @xmath200 for ( a ) @xmath117 and ( b ) @xmath119 .
here we have used @xmath203 gev and @xmath206 . ]
the projection of the difference between @xmath208 and @xmath209 events presents a clear signature in the mass spectrum with a zero close to the mass of the vector meson @xmath1 , and another one close to the mass of the scalar @xmath2 .
there is also a change of sign in the cp asymmetry associated to the @xmath28 passing through zero around the middle of the dalitz plot .
these features can be observed in fig .
[ dcpv - f ] , where we display the term containing the @xmath207 parameter of the integrated eq .
( [ dgamacos ] ) .
this term locally satisfies the cpt constraint .
parameter of the integrated eq .
( [ dgamacos ] ) , as a function of @xmath200 for ( a ) @xmath117 and ( b ) @xmath119 . here
we set @xmath210 . ]
the shape of this term presents the characteristic peaks of the two interfering resonances . due to the direct proportionality to @xmath212
, there is a change of sign of the cp asymmetry when @xmath28 pass through zero in the middle of the dalitz plot .
[ dcpv - g ] shows the features of this term .
this term is consistent with the cpt constraint locally .
parameter of the integrated eq .
( [ dgamacos ] ) , as a function of @xmath200 for ( a ) @xmath117 and ( b ) @xmath119 . here
we set @xmath213 . ]
we determine the relative contributions related to the integrated eq .
( [ dgamacos ] ) by fitting two cp asymmetries distributions presented in ref .
@xcite ; one for the @xmath22 mass distribution coming from the difference between @xmath208 and @xmath209 in @xmath19 decay , in particular the distribution with @xmath119 . from this fitting
we found the parameters and plot the cp asymmetry for @xmath117 .
the other fit was performed for the phase - space integration in the region @xmath214 of the @xmath22 mass distributions in the @xmath215 decay . from these two fits
we get the value of the @xmath216 parameter @xmath217 for each decay channel and plot the asymmetry for the @xmath23 distribution in the channels @xmath6 and @xmath21 . to perform both fits and the other plots , we use the @xmath216 amplitude , the non resonant component , and the @xmath1 and @xmath2 resonances .
we put these amplitudes within the isobar model through a coherent sum of them .
the biggest source of uncertainty from these fits is the error in the @xmath7 phase shift and inelasticity parameter .
for simplicity we use the central value of ref .
@xcite , although possible variations within the quoted errors in the parameters could give better results .
despite our belief that this model is able to represent much of this data , we are aware that it can not explain all the rich cp violation structure observed in these decays
. the inclusion of other contributions and the symmetrization of the @xmath218 decay amplitude would be necessary to understand in more details the phase space of these decays and get better agreement in all cp asymmetry regions presented in ref .
@xcite . to perform the fits we use for the @xmath2 resonance , @xmath170 gev and width @xmath173 gev , getting from e791 experiment @xcite .
for the two @xmath219 parameters , we use @xmath220 gev and @xmath221 gev .
it is important to say that we changed these values by a factor two without finding an appreciable change in the fitting .
we started by performing the fit of the integrated eq .
( [ dgamacos ] ) to the @xmath119 asymmetry distribution of the low @xmath22 mass projection for the @xmath19 decay . due the presence of identical particles in final state
, the interference between symmetrical terms by the exchange of the two identical pions must disturb the cp violation pattern mostly for @xmath117 .
we study the regions on the dalitz plot where this interference can be minimum and observe that it corresponds to the high @xmath223 region on the @xmath224 distribution as can be seen in fig .
[ m32 ] .
this figure provides the mass of the changeless pair of pions formed with the bachelor pion .
if @xmath117 the mass of this pair stays below 3 gev , and for @xmath225 gev , it can be even below 1 gev , and therefore making relevant the interference with the resonances . for @xmath119 , @xmath226 gev , minimizing interference effects from the bose symmetrization of the decay amplitude .
we intend to perform this study in the future . for the moment
, we do not consider the symmetrization of the decay amplitude , and therefore we do not use the data for the cp asymmetry in the @xmath19 channel for @xmath117 to fit the parameters .
the best fit was obtained by using four parameters associated to the @xmath227 amplitude ( @xmath193 ) , the real and imaginary parts of the interference between the @xmath1 and the non resonant partonic amplitudes ( @xmath202 and @xmath205 ) , and finally the imaginary part of the interference between the @xmath1 and the @xmath2 resonances ( @xmath211 ) . the result is presented in fig .
[ b - pipipi - cos - pos]a , where we compared our model with the experimental data extracted from fig .
4c of ref .
the remaining parameters , namely , those that locally violate the cpt constraint , and those negligible in the fitting procedure , were set to zero .
[ b][b][0.8]@xmath29 [ b][b][0.8]@xmath119 [ b][b][0.8]@xmath193 term [ b][b][0.8]@xmath202 term [ b][b][0.8]@xmath205 term [ b][b][0.8]@xmath211 term decay , integrated eq .
( [ dgamacos ] ) , compared with the experimental values ( blue points ) taken from ref . @xcite . results for @xmath119 for ( a ) total and ( b ) individual contributions.,title="fig : " ] the individual contributions to the cp asymmetry are shown in fig . [ b - pipipi - cos - pos]b .
as expected , the @xmath1 meson contribution represented by the amplitudes containing the @xmath202 and @xmath205 parameters , is large mainly for the real part of the bw , and the presence of @xmath2 is seen only by one of the interfering terms with the @xmath1 , namely , that presenting the @xmath211 parameter .
all these terms are locally cpt invariant , namely by integration in @xmath114 their contribution to the cp asymmetry vanishes . between @xmath12 and @xmath8 gev
, the contribution of the @xmath216 amplitude , namely , the @xmath193 term , shows its importance and as expected @xcite , dominates the asymmetry in this region .
we remind that this part of the asymmetry does not vanishes upon integration in @xmath28 and cancels the asymmetry in the @xmath6 decay channel .
looking at the behavior of the experimental @xmath214 distribution in fig .
[ b - pipipi - cos - neg]a , we can see that much of the features observed for the distribution of events for @xmath224 are present in this asymmetry .
in fact , we plotted together with the experimental points , the amplitudes computed with the integrated eq .
( [ dgamacos ] ) using the fit parameters obtained from the @xmath119 distribution .
the plot in fig .
[ b - pipipi - cos - neg]a shows a clear departure from the experimental data when at the starting of the @xmath1 mass resonance and also at the beginning of the contribution of the @xmath216 amplitude to the asymmetry .
assuming that these differences are due to the symmetrization , this result suggests an interference between @xmath1 and @xmath216 amplitude in the crossing channels , which is corroborated by fig .
[ m32 ] , where for the mass region above @xmath12 gev , the mass of the @xmath228 pair in the crossing channel for @xmath229 can be even below @xmath12 gev .
in addition , we show in fig .
[ b - pipipi - cos - neg]b the individual contributions to this cp asymmetry distribution .
[ b][b][0.8]@xmath29 [ b][b][0.8]@xmath117 [ b][b][0.8]@xmath193 term [ b][b][0.8]@xmath202 term [ b][b][0.8]@xmath205 term [ b][b][0.8]@xmath211 term decay , integrated eq .
( [ dgamacos ] ) , compared with the experimental values ( blue points ) taken from fig .
4d of ref .
results for @xmath117 for ( a ) total and ( b ) individual contributions.,title="fig : " ] we still remark to the reader that for the @xmath19 decay , our definition of the angle @xmath96 is opposite from that presented in ref .
thus , we had to compare our results of @xmath119 ( @xmath117 ) with those of @xmath117 ( @xmath119 ) of ref .
@xcite .
the expression for the cp asymmetry in the coupled @xmath30 channel , derived from the general formula in eq .
( [ cp27b ] ) , applied to this specific decay and integrated in @xmath28 , is given by @xmath230}{a'(s)\sqrt{s-4m_k^2 } \left(1 + \frac{s}{\lambda_{\lambda}^2}\right ) \left(1 + \frac{s}{\lambda_{\lambda^\prime}^2}\right ) } , \label{dgpikk } \end{aligned}\ ] ] where the kinematical factors , namely , @xmath231^{1/2 } } , \label{aprime}\end{aligned}\ ] ] and the kaon momentum in the rest frame of the @xmath100 subsystem , @xmath232 , are now taken for the @xmath137 system .
furthermore , the integrated decay width in @xmath28 from eq .
( [ dgpikk ] ) , becomes exactly opposite in sign to the decay width in the @xmath222 channel above the @xmath100 threshold , c.f .
the first two terms of the integrated eq .
( [ dgamacos ] ) . the expression in eq .
( [ dgpikk ] ) is obtained by adding the @xmath216 contributions detailed in figs .
[ b - pipipi - cos - pos ] and [ b - pipipi - cos - neg ] , since @xmath233 .
the plot shown in fig .
[ b - pikk ] was done with the parameters fitted by the cp asymmetry data in the decay .
the cp violation distribution obtained for the @xmath23 invariant mass from @xmath6 decay through eq .
( [ dgpikk ] ) , has opposite sign with respect to the correspondent term in the coupled @xmath165 channel .
[ b][b][0.8]@xmath30 ) , compared with experimental data ( blue points ) taken from fig .
7b of ref . @xcite.,title="fig : " ] for this decay ,
the lhcb experiment presented the sum of events with @xmath117 and @xmath119 .
the comparison of the data with our model is also provided in fig .
[ b - pikk ] .
there is a clear agreement about the shape of the model distribution with the experimental data and also a reasonable amount of the number of events related with this kind of cp asymmetry .
it is also worth to mention that cpt is conserved if we sum all cp asymmetry contributions obtained with our approach to @xmath218 and @xmath6 decays , in a region of @xmath22 and @xmath23 invariant mass below @xmath8 gev .
the number of events observed in these two decay channels are about one order of magnitude larger than in the @xmath165 and @xmath167 data @xcite .
this allows a better fit to the difference between @xmath208 and @xmath209 events . in the fitting procedure
we use only the @xmath117 distribution , because the experimental results @xcite for @xmath119 present a new feature in both decays in the region studied in this work .
our simple model does not account for this new behavior .
we only have a guess , that we already mentioned in our previous paper @xcite , related with the possible presence of re - scattering coming from double charm decays .
one should note that three light - pseudoscalar mesons can , in principle , couple via strong interaction with channels like @xmath234 .
it seems reasonable to expect that @xmath235 or @xmath236 can contribute to the cp asymmetry in regions of large two - body invariant mass above the @xmath237 threshold , that is far from the @xmath100 threshold and above @xmath8 gev , outside the region discussed in this work but excluded from the @xmath117 distribution . differently to the @xmath19 decay where the @xmath1 amplitudes were dominant , here the @xmath2 has the largest contribution to the @xmath26 decay , as it is expected by recalling the bss mechanism applied to these decays , which builds in principle the corresponding amplitudes associated with the source terms or partonic amplitudes .
one possible good fit for the integrated eq .
( [ dgamacos ] ) is obtained by using five parameters associated to the @xmath238 amplitude ( @xmath193 ) , the real parts of the interference between @xmath1 resonance with the non resonant partonic ( @xmath202 ) , and @xmath239 amplitudes ( @xmath240 ) , and finally the real and imaginary parts of the interference between the @xmath1 and the @xmath2 resonances , ( @xmath207 and @xmath211 ) . as in the @xmath19 case ,
the parameters that locally violate the cpt constraint , and those negligible in the fitting procedure were set to zero .
[ b - kpipi - cos - neg]a shown the best fit we get to the @xmath214 distribution .
[ b][b][0.8]@xmath31 [ b][b][0.8]@xmath117 [ b][b][0.8]@xmath193 term [ b][b][0.8]@xmath202 term [ b][b][0.8]@xmath207 term [ b][b][0.8]@xmath211 term [ b][b][0.8]@xmath241 term decay , integrated eq .
( [ dgamacos ] ) , compared with the experimental values ( blue points ) taken from fig .
5c of ref .
results for @xmath117 for ( a ) total and ( b ) individual contributions.,title="fig : " ] in fig .
[ b - kpipi - cos - neg]b , we shown the decomposition of each component of the fit .
clearly the dominant contribution comes from the real part of the interference between @xmath1 and the @xmath2 resonances .
the imaginary part of this amplitude has a contribution , however , much less important than the real one .
the @xmath22 and @xmath242 amplitude plays an important role in this fit above @xmath12 gev and also associated with the interference with the @xmath1 resonance . here
, we also used @xmath243 gev and @xmath244 gev .
the other parameters were found by using the @xmath245 method in order to fit the experimental data for @xmath117 distribution from ref .
@xcite .
the cp asymmetry in the @xmath208 and @xmath209 event distributions in the @xmath21 channel is compared to our model with the sum of the events of @xmath117 and @xmath119 as we did for @xmath6 decay channel and the data provided by the lhcb in this last case .
the functional form of the asymmetry for the @xmath27 decay is the same as in eq .
( [ dgpikk ] ) .
the only change is due to the kinematical factor @xmath246 in eq .
( [ aprime ] ) , that now presents the replacement @xmath247 .
we plot in fig .
[ b - kkk ] the asymmetry of the @xmath27 decay , computed with the parameter @xmath193 obtained in our previous fitting of the decay . our model is compared to the data obtained by summing the experimental results of @xmath27 distributions for both @xmath28 positive and negative regions .
we can go beyond the asymmetry formula in eq .
( [ cp27d ] ) , where we have not included the interference terms between the resonances and the scattering terms in the elastic channel , we remind that in the bw we included the elastic amplitude , but it is also important to consider it outside the resonance region , and returning to ( [ cp27b ] ) as the starting point . we have neglected the bose symmetrization of the amplitude by the exchange of the identical pions and the contribution of the double charm production and the its coupling to the @xmath100 channel .
this also should be addressed in the future .
the isoscalar and zero angular momentum @xmath7 originated by contribution of the elastic scattering terms to the cp asymmetry for and decays written before , can be substituted the inelasticity and phase - shifts given by @xmath144 s - matrix provided by eq .
( [ s - matrix ] ) as @xmath248 .
the inclusion of the @xmath249wave @xmath7 scattering should improve the fitting below the @xmath1 , where it is clearly seen in fig .
[ b - pipipi - cos - neg ] , the effect of a missing interference term . in this particular region
the @xmath250 , namely the @xmath251 , will be taken into account by @xmath252 , and we expect to improve the fit in the low mass region .
the interference between @xmath253 and the @xmath1 resonance was used to fit the @xmath254 channel , as we can see in fig .
[ b - kpipi - cos - neg ] .
other possibility in this fit would be consider the large presence of @xmath2 , by using its interference with @xmath253
. this component do not satisfy cpt locally , but is consistent with eq .
( [ cp27b ] ) with the coupled channel @xmath255 . in this work ,
we choose to use only locally cpt invariant components for the interference with resonances .
we show how cp asymmetry in charmless three - body @xmath4 decays is constrained by cpt in the presence of resonances and final state interaction in leading order .
the cpt constraint @xmath256 , should include in the sum over the final states , all the kinematical allowed phase space and possible decay channels . in leading order , the scattering matrix @xmath257 corresponding to the decay channel where a resonance is formed , e.g. , the @xmath1 in p - wave @xmath7 elastic amplitude , accounts for the dressing of the decay vertex @xmath258 .
the physical decay amplitudes associated with the partonic source , that give the @xmath1 meson and in its sequential decay to two pions , is dressed by the @xmath7 re - scattering process in the resonant channel .
the non - resonant channels receive contribution from the re - scattering in leading order .
the decay of the resonance and re - scattering by the bachelor meson , should appear as a next - to - leading order contribution , to be consider in future works , where the general formulas for the three - body decay amplitude based on the cpt constraint that includes the formation of resonances , re - scattering and coupling to inelastic channels should be addressed . in our limited two - body description
, we explored the rich patterns of the cp asymmetry formula , namely interference between non - resonant and resonant terms , as well as the coupling between the @xmath7 and @xmath100 channels due to the strong interaction .
we analysed in our study two examples of cp violation in charmless
@xmath4 three - body decays in the presence of resonances , excluding the possibility of channel coupling .
the first case discussed is the @xmath72 meson resonance plus a non resonant channel , and the second one is the case of two interfering resonances , namely , @xmath72 and @xmath2 .
these examples discussed within two amplitude models are suitable for the study of @xmath19 and @xmath20 for @xmath259 masses below the @xmath100 threshold , as provided by the new data lhcb data @xcite , and we selected terms that are locally constrained by cpt .
the cp asymmetry produced by the @xmath72 meson resonance in the simple two amplitude model with a non - resonant channel , which are brought by the real and imaginary parts of the interference term , presents a clear signature in the low mass @xmath260 spectrum for both decays @xmath19 and @xmath20 , as indeed the new data suggests @xcite .
one term shows a very clear signature : a zero at the @xmath72 mass in both regions of @xmath117 and @xmath261 for @xmath19 with a change of the sign around @xmath262 .
the two amplitude model constituted by the resonances @xmath72 and @xmath2 , shows two zeros in the cp asymmetry from the real part of the breit - wigners : one close to the @xmath72 mass and another one close to the @xmath2 mass .
however , the term which mixes the real and imaginary parts of the breit - wigners does not present a clear signature of cp violation around the @xmath72 meson mass , while a zero is present at the @xmath2 mass , because the ratio @xmath263 is considerably larger than one . in both examples ,
the zero s of the asymmetry at the resonance mass are associated only with the dalitz cp violation from the interference of two amplitudes and the real part of the breit - wigner , the contributions associated with the imaginary part present a behavior with a maximum at the resonances masses .
the recent experimental results @xcite are strongly indicating that the cp asymmetry appearing in the low @xmath260 mass region comes from the real part of the breit - wigners in the interference between two amplitudes involving the @xmath1 meson resonance .
in addition , the cp asymmetries in whole kinematical region below @xmath264 gev in @xmath19 and @xmath26 decays demand the coupling to the @xmath265 by the strong interaction , as has already been discussed in ref .
@xcite . to account for that
, we introduced in the fitting of the cpv , the contribution of the coupled @xmath7 to @xmath100 channels in isoscalar and zero angular states .
this was relevant for the fitting above the @xmath100 threshold .
the independent confirmation of the cp asymmetry flowing between different channels coupled by the strong interaction was checked by the fair reproduction of the cpv in @xmath30 and @xmath10 distributions .
the computation of the asymmetry was performed as dictated by the cpt constraint applied to coupled channels .
there was no new parameters beyond those given in the fitting of the cp asymmetry data in the @xmath266 and @xmath20 decays .
we discussed improvements to cp asymmetry formula beyond the one used to fit the experimental data on decay , where the interference between the resonances and elastic scattering terms are missing .
the inclusion of the interference between the @xmath249wave @xmath7 scattering amplitude and the @xmath72 meson breit - wigner should improve the fit below the @xmath267 . in this particular region , the @xmath250 , namely the @xmath251 , will be taken into account through the elastic @xmath7 scattering , which it is expected to improve the fit .
still the bose symmetrization of the amplitude by the exchange of the identical pions has to be considered , which should be addressed in the future .
finally , the results of our fittings suggest an important consequence to cp violation : the strong phases can come from the hadronic interaction minimizing the effect of the imaginary part of penguin contributions to both decays studied here , as the recent experimental results are indicating .
by defining in fig .
[ b - pipipi - theta ] , @xmath268 , @xmath269 , and @xmath270 ( bachelor ) , one has that @xmath271 , and @xmath272 .
the kinematics of the decay imposes that @xmath28 is a function of @xmath273 and @xmath274 , namely , @xmath275 where , @xmath276^{1/2}},\end{aligned}\ ] ] and @xmath277^{1/2}},\quad\end{aligned}\ ] ] with @xmath278 gev ( the detailed derivation of this equation is relegated to the appendix [ app - cos ] ) . therefore , the asymmetry presented in eq .
( [ dgamacos ] ) is actually a function of @xmath273 and @xmath274 , due to @xmath28 . in order to eliminate @xmath279
, we integrate @xmath280 in this variable , leading to the asymmetry @xmath281 .
we still separate the integration in two regions , namely , that of @xmath117 , leading to @xmath282 and that of @xmath119 , leading to @xmath283 the quantities @xmath284 correspond , respectively , to @xmath285 and @xmath286 , according to eq .
( [ costheta ] ) .
the quantity @xmath287 is related to @xmath288 . the integration in eqs .
( [ dg - cos - neg ] ) and ( [ dg - cos - pos ] ) generates the following asymmetry , @xmath289\sqrt{1-\eta^2(s)}}{a(s)\sqrt{s-4m_\pi^2 } \left(1+\frac{s}{\lambda^2_{\lambda}}\right)\left(1+\frac{s}{\lambda^2_{\lambda^\prime } } \right ) } % + \frac{\mathcal{c}|f^{\rm
bw}_\rho(s)|^2k^2(s)}{3a(s ) } \nonumber \\ % & + & \frac{|f^{\rm bw}_\rho(s)|^2k(s)}{2\xi a(s ) } \left\{\frac{\mathcal{d}(m^2_\rho - s)}{1+\frac{s } { \lambda^2_{\lambda } } } + \frac{\mathcal{d}^\prime\sqrt{1-\eta^2(s)}\left\{m_\rho\gamma_\rho(s)\cos[2\delta_{\pi\pi } ( s)]-(m^2_\rho - s)\sin[2\delta_{\pi\pi}(s)]\right\}}{\sqrt{s-4m_\pi^2 } \left(1+\frac{s}{\lambda^2_{\lambda^\prime}}\right ) } \right .
% & + & \left .
\frac{\mathcal{e}m_\rho\gamma_\rho(s)}{1+\frac{s } { \lambda^2_{\lambda } } } + \frac{\mathcal{e}^\prime\sqrt{1-\eta^2(s)}\left\{(m^2_\rho - s)\cos[2\delta_{\pi\pi } ( s)]+m_\rho\gamma_\rho(s)\sin[2\delta_{\pi\pi}(s)]\right\}}{\sqrt{s-4m_\pi^2 } \left(1+\frac{s}{\lambda^2_{\lambda^\prime}}\right)}\right\ } \nonumber \\ % & + & \frac{|f^{\rm bw}_{\rho}(s)|^2|f^{\rm bw}_f(s)|^2k(s)}{2\xi a(s)}\times \nonumber\\ & \times&\left\{\mathcal{f}[(m_\rho^2 -s)(m_f^2-s ) + m_\rho\gamma_\rho(s)m_f\gamma_f(s ) ] + \mathcal{g}[(m_\rho^2 -s)m_f\gamma_f(s ) - m_\rho\gamma_\rho(s)(m_f^2-s)]\right\ } \nonumber \\ % & + & \frac{|f^{\rm bw}_f(s)|^2}{a(s ) } \left\{\frac{\mathcal{h}(m^2_f - s)}{1+\frac{s}{\lambda^2_{\lambda } } } + \frac{\mathcal{h}^\prime\sqrt{1-\eta^2(s)}\left\{m_f\gamma_f(s)\cos[2\delta_{\pi\pi } ( s)]-(m^2_f - s)\sin[2\delta_{\pi\pi}(s)]\right\}}{\sqrt{s-4m_\pi^2 } \left(1+\frac{s}{\lambda^2_{\lambda^\prime}}\right ) } \right .
% & + & \left .
\frac{\mathcal{p}m_f\gamma_f(s)}{1+\frac{s } { \lambda^2_{\lambda } } } + \frac{\mathcal{p}^\prime\sqrt{1-\eta^2(s)}\left\{(m^2_f - s)\cos[2\delta_{\pi\pi } ( s)]+m_f\gamma_f(s)\sin[2\delta_{\pi\pi}(s)]\right\}}{\sqrt{s-4m_\pi^2 } \left(1+\frac{s}{\lambda^2_{\lambda^\prime}}\right)}\right\ } + \frac{\mathcal{q}|f^{\rm bw}_f(s)|^2}{a(s ) } , \label{dgama - final } \end{aligned}\ ] ] where @xmath290 for @xmath117 , and @xmath291 for @xmath119 . here , the rest frame momentum of the pion , namely , @xmath292 , is included in the terms presenting interference with the @xmath253 amplitude , in order to introduce the kinematical factor in the scattering amplitude . furthermore
, this factor is consistent with the cpt constraint , which indicates that the decay widths for the coupled @xmath222 and @xmath137 channels above the @xmath100 threshold is such that @xmath293 , as discussed at the end of sec .
[ sec : pikk ] .
finally , we also remind that all free parameters are compatible with the energy quantities of the model , that are in gev units .
in a general way , one has that @xmath294 if @xmath295 , then eq . ( [ ap1 ] ) becomes @xmath296 in the decay of fig .
[ b - pipipi - theta ] , we define @xmath268 , @xmath269 , and @xmath270 ( bachelor ) . therefore , @xmath271 , @xmath272 , and @xmath297 , since we have degenerated the pion mass .
if we assume that @xmath298 then , these definitions lead to @xmath299 thus , @xmath28 is written as @xmath300 in order to write @xmath28 in terms of @xmath279 and @xmath301 , it is necessary to find @xmath302 , @xmath303 , @xmath304 , and @xmath305 .
the first two quantities can be extracted as follows .
notice that , @xmath306 furthermore , by the figure , we see that @xmath307 .
however , since @xmath308 and @xmath309 , we conclude that @xmath310 . then
, @xmath311 with @xmath312 and @xmath313 in order to find the @xmath273 dependence of @xmath304 and @xmath305 , we proceed to write @xmath314 and make @xmath315 in the reference frame in which @xmath316 , we have @xmath317 furthermore , in this reference frame , @xmath318 , leading to @xmath319 from this equation we get , @xmath320 finally , from @xmath321 , we have @xmath322 inserting eqs .
( [ ppi ] ) , ( [ p0pi ] ) , ( [ p03 ] ) and ( [ p3 ] ) in eq . ( [ costheta - ap ] )
we get , @xmath323^{1/2}}\nonumber \\ % & = & \frac{2m_{32}^2 - m_b^2 - 3m_\pi^2 + s}{2\left(s-4m_\pi^2\right)^{1/2 } \left[\frac{(m_b^2-m_\pi^2-s)^2}{4s}-m_\pi^2\right]^{1/2}}.\qquad \label{costhetafinal - ap}\end{aligned}\ ] ] this same equation is obtained if we now treat the decay . in this case , the @xmath0 meson and the pions of fig .
[ b - pipipi - theta ] have its charges changed , and one defines @xmath324 , @xmath325 and @xmath326 ( bachelor ) .
therefore , one has @xmath327 , and @xmath328 . for this case
, @xmath28 will be written exactly as in eq .
( [ costhetafinal - ap ] ) , but for @xmath329 .
we thank the support from conselho nacional de desenvolvimento cientfico e tecnolgico ( cnpq ) of brazil .
o. l. also acknowledges the support of the grant @xmath3302013/26258 - 4 from so paulo research foundation ( fapesp ) .
j. h. a. n. acknowledges the support of the grant @xmath3302014/19094 - 8 from so paulo research foundation ( fapesp ) . | resonances and final state interactions ( fsi ) play a role in the formation of cp violation ( cpv ) constrained by cpt invariance .
we provide a general formulation of cpv including resonances and fsi starting from the cpt constraint .
our discussion is elaborated within a simple @xmath0 decay model with the @xmath1 and @xmath2 resonances plus a non resonant background including the @xmath3 coupled amplitude .
we consider few illustrative examples to show the interference patterns appearing in the cp asymmetry , namely , that from the @xmath1 resonance plus a non - resonant amplitude , and that from the interference of the @xmath1 and @xmath2 resonances .
we perform the fit of the cp asymmetry for the charmless three - body @xmath4 decay channel @xmath5 and obtain as outcome the @xmath6 for @xmath7 channel asymmetry in the mass region below @xmath8 gev in fair agreement with the new data lhcb data .
analogously , we also describe the cp asymmetry of the @xmath9 decay , with that from the @xmath10 channel obtained as output . as in the previous case
, we also found agreement with lhcb experimental data . |
presently , modern surgery has at its disposal a great variety of different surgical techniques to heat biological tissue in a localized and safe way .
all these techniques are based on specific applicators , _
i.e. _ devices which extract ( cryosurgery ) or introduce heat ( laser , radiofrequency current , microwave or ultrasound treatments ) .
theoretical heat modeling is a very cheap and fast methodology to study the thermal performance of these applicators .
in fact , a lot of previous work has been conducted to model these procedures by using the bioheat equation as governing equation ( see @xcite , and references therein ) .
this equation is based on the classical fourier theory of heat conduction and is widely used for modeling the heating of biological tissue .
nevertheless , it has been hypothesized that for heat transfers on very small time scales the classical model will fail , and an alternative thermal wave theory with a finite thermal propagation speed could be more suitable to describe these phenomena . for the aforementioned reasons
, we are lead to modeling the surgical heating of biological tissue by means of the hyperbolic heat transfer equation , and then compare these alternative results with those obtained from the standard fourier theory ( implying the parabolic heat transfer equation ) . in summary ,
the main theme of this work is to present in depth the mathematical and physical background for a general discussion of heat models related to heating of biological tissue by means of energy applicators , and then , in particular , point out the physical differences between the classical fourier theory and the hyperbolic wave theory .
moreover , as an illustrative example , we give some numerical results of the analytical modeling for radiofrequency heating ( _ rfh _ ) and for laser heating , each applied to the cornea in order to correct refractive errors .
both surgical techniques have in common that they may involve high heat transfers during very short exposure times .
this paper is organized as follows . in sec .
[ heatmodels ] we will set up the mathematical groundwork for heat models related to _
rfh _ and laser heating with a special emphasis on parabolic and hyperbolic heat models .
then , in sec . [ applications ] we present the results for typical laser and _ rfh _ interventions in order to substantiate the physical discrepancies between both heat models in concrete examples .
we will also comment on the domain of applicability of both theories and their main features .
finally , sec .
[ conclusions ] will present the conclusions and give an outlook on interesting future work .
the most fundamental relation to model heat transfer is the generally valid , _
i.e. _ model - independent , equation for thermal current conservation with a given internal heat source @xmath0 : @xmath1 here , @xmath2 denotes the thermal flux and @xmath3 is the temperature at point @xmath4 in the domain @xmath5 at time @xmath6 . as usual ,
thermal conductivity is denoted by @xmath7 and diffusivity by @xmath8 , where @xmath9 is the volumetric heat capacity , being @xmath10 the density and @xmath11 the specific heat of the material under consideration . fig .
[ currentpic ] gives a schematic view of thermal current conservation .
fourier s law of heat conduction @xcite constitutes the foundation of all classical heat models .
it proposes that the heat flux @xmath2 is proportional to the negative of the temperature gradient ( see fig .
[ fourierpic ] ) : @xmath12 however , in this relation temporal changes instantaneously affect heat flux and temperature gradient . as an immediate consequence
any perturbations in classical heat models are propagated with infinite speed . assuming that thermal current conservation ( [ current ] ) and fourier s law ( [ fourier ] ) hold , directly yields the classical , _
i.e. _ parabolic heat transfer equation ( _ phte _ ) with heat sources : @xmath13 the standard bioheat equation was introduced by pennes @xcite and is based on the parabolic heat transfer model , identifying the following explicit contributions for the heat sources in a biological system @xmath14 where the subindex @xmath15 denotes a surgical heat source ( _ e.g. _ laser or radiofrequency treatment ) , @xmath16 refers to blood perfusion , and @xmath17 to any source related to metabolic activity . with the advancement of modern surgery ,
medical treatments involve progressively smaller time scales and higher energy fluxes .
for example , for radiofrequency and laser surgery _
phte _ models could become inappropriate .
the reason is that on small time scales @xmath18 $ ] , with a sufficiently small value for @xmath19 , thermal equilibrium of an extended physical system simply can not be reached .
this contradicts the classical model , where all elements of the thermodynamic system interact with no delay through infinite thermal speed .
it is to be expected that such theories implying infinite speed of perturbations in the heated medium and more realistic wave theories with finite speeds differ considerably in their predictions .
a modified version of fourier s law ( [ fourier ] ) uses a non - vanishing relaxation time @xmath20 in the dissipative process .
this parameter may be interpreted as the finite time necessary for the dissipative flow to relax to its steady thermodynamic value .
the simplest generalization for finite speeds leads to cattaneo - vernotte s equation @xcite : @xmath21 note that the classical fourier law is still included as a special limiting case , taking the limit of zero relaxation time for the heat fluxes . in general
, however , there is an inertial or retardation term which delays any changes in the temperature gradient to be transferred to the heat flux .
this can also be seen for small relaxation intervals @xmath20 , which yields the expression @xmath22 although it appears at first sight that eq .
( [ cv ] ) is an expansion which only holds approximately for small relaxation intervals @xmath20 , by considering entropy arguments ziik and tzou have shown that eq .
( [ cv ] ) is also valid on a macroscopic scale @xcite . by combining the heat transfer model ( [ cv ] ) with current conservation ( [ current ] ) , one directly obtains the hyperbolic heat transfer equation ( _ hhte _ ) with heat sources : @xmath23 for zero relaxation the hyperbolic heat equation obviously reduces to the classical model . an essential feature of eq .
( [ hhte ] ) is the second - order time derivative with coefficient @xmath19 , which gives rise to wave properties of the equation , and therefore implies finite thermal propagation speeds .
this can easily be seen by considering forwardly propagating waves which provide solutions for eq .
( [ hhte ] ) of the form @xmath24 where @xmath25 is the usual wave propagation vector and @xmath26 the corresponding frequency .
this approach is somehow analogous to the analysis of the _ telegraph equation _ , which describes the propagation of electromagnetic waves in conducting media @xcite .
note that in the present case , however , a term proportional to @xmath27 is absent on the left - hand side of eq .
( [ hhte ] ) . by following the procedure of ref .
@xcite and substituting eq .
( [ wave ] ) into eq .
( [ hhte ] ) for sourceless heat transfer , then requiring that the real - numbered factor multiplying with @xmath28 has to be zero , one directly obtains for the thermal propagation speed @xmath29 from this it becomes evident that for vanishing relaxation times @xmath20 and constant diffusivity @xmath30 , the propagation speed tends to infinity . as indicated before , but seen from a different perspective , in this limiting case the hyperbolic model reduces again to the standard fourier theory .
one further distinctive feature of the _ hhte _ model is observed by representing the hyperbolic heat flux equation ( [ cv ] ) in integral form @xcite @xmath31 note that eq .
( [ hypflux ] ) shows that @xmath32 depends on the full history of the temperature gradient within the time interval @xmath33 $ ] , something which is entirely absent in the much simpler expression for the fourier flux in eq .
( [ fourier ] ) . exhibiting these features ,
the hyperbolic heat transfer model with its finite thermal propagation appears to provide an ideal framework to describe surgical interventions such as laser ablation ( a procedure named as laser thermokeratoplasty , or short _ ltk _ ) and _ rfh _ of the cornea ( _ ck _ , conductive keratoplasty ) , which use very short and high - energetic pulsations ( typical time periods are @xmath34 and @xmath35 , respectively ) .
therefore , in the following section , we will focus on applications of heat models in surgical techniques , in particular _
rfh _ and laser surgery applied to the cornea , and thereby develop a realistic model for the heating of biological tissue by employing the _ hhte _ as previously derived in eq .
( [ hhte ] ) .
there exist numerous processes in which great amounts of heat are applied to materials in very short exposure times .
this section will focus on the modeling of heating biological tissue as it occurs during medical intervention with surgical radiofrequency or laser devices to the cornea .
laser heating is a process which implies tissue heating caused by absorbing the optical energy of a high - energy laser beam ( _ ltk _ ) . on the other hand , radiofrequency heating ( _ rfh _ )
is ultimately based on ohm s law , which states that currents flowing through a resistor generate heat .
it is essentially ionic motion in the tissue which will provoke biological heating aimed at producing any medical effects .
both are surgical techniques which may not only involve small time scales but also high energy fluxes , which makes them an important group of applications in which differences between parabolic and hyperbolic models could have great effects . as we have previously mentioned in sec .
[ heatmodels ] , these type of processes represent non - equilibrium processes on small time scales ( _ i.e. _ compared to the relaxation time @xmath20 ) , since the system requires considerably more time to reach the equilibrium state after the initial thermal energy input . for this reason , _ phte _ models , with their infinite thermal propagation speed , may not provide an appropriate description for the underlying physical structure , and it may be necessary to rely on _ hhte _ models instead . among the many surgical procedures in which laser heating is employed , we focus our attention on corneal laser heating , also referred to as laser thermokeratoplasty ( _ ltk _ )
@xcite . in the case of laser heating
, we consider a theoretical model consisting of a semi - infinite fragment of homogeneous isotropic biological tissue in which the laser beam falls on the entire tissue surface .
[ laserpic ] depicts the chosen model geometry , in which we consider a one - dimensional model with the @xmath36-axis parallel to the direction of the incident laser beam . above the surface , for @xmath37 , free thermal convection
is supposed to cool the tissue surface .
as we want to study the problem from the point of view of the _ hhte _ model , we have to start off from the governing equation provided by eq .
( [ hhte ] ) . for this one - dimensional laser heating model
, we have to solve the corresponding heat transfer equation for penetration depth @xmath36 including the appropriate heat - source term and boundary conditions . neglecting blood perfusion and metabolic activity in eq .
( [ bioheat ] ) , and hence @xmath38 , the heat source @xmath39 will only have contributions from @xmath40 .
this surgical source should be obtained from the beer - lambert law , which empirically states that for radiation the intensity decreases exponentially with penetration depth . including a factor with the temporal dependence to model an energy pulse of duration @xmath41 , this yields @xmath42,\ ] ] where @xmath43 denotes the dimensionless fresnel surface reflectance , @xmath44 is the absorption coefficient , and @xmath45 is the incident energy flux at the tissue surface @xcite . as usual
, @xmath46 denotes the heaviside function . after combining eqs .
( [ hhte ] ) and ( [ beer ] ) , the final form of the governing equation is @xmath47 } , \end{array}\ ] ] where @xmath48 is the dirac delta function .
the corresponding initial boundary conditions are given by @xmath49 \lim\limits_{x\to\infty } t(x , t)=t_0 & & \qquad\forall t>0 \\[.25 cm ] \displaystyle{\partial t\over\partial x}(0,t ) = \displaystyle { { h\over k}\left(\tau \frac{\partial t}{\partial t}(0,t)+t(0,t)-t_a\right ) } & & \qquad\forall t>0 \label{coolflux } \end{array}\ ] ] where @xmath50 is the initial temperature and @xmath51 is the ambient temperature .
furthermore , @xmath52 is the thermal convection constant at the interface tissue given by newton s law of cooling , which states that the convective flux of an object is proportional to the difference between its own temperature @xmath50 and the ambient temperature @xmath51 : @xmath53.\ ] ] since at the interface tissue with @xmath54 the heat flux is given by the newton s law of cooling , the last condition of eq .
( [ coolflux ] ) is obtained by imposing this law onto the relation for the hyperbolic heat flux eq .
( [ hypflux ] ) ( see also @xcite ) .
the explicit analytical solution of this problem has been obtained in ref .
@xcite and is mainly based on the use of laplace transforms
. it also contains for comparison the fully analytical _
solution as an application to the thermokeratoplasty technique . for the numerical estimates ,
well - established thermal and optical properties of the cornea were used .
all essential physical constants for the model are summarized in tab .
[ corneaparam ] .
[ corneaparam ] density & conductivity & diffusivity & convection + @xmath10 [ kg m@xmath55 & @xmath56 [ w m@xmath57 k@xmath57 ] & @xmath59 [ m@xmath60 s@xmath57 ] & @xmath61 [ w m@xmath62 k@xmath57 ] + @xmath63 & @xmath64 & @xmath65 & 20 + fresnel reflectance & absorption coefficient + @xmath43 & @xmath44 [ m@xmath57 ] + 0.024 & 2000 + .35 cm figure [ laserabl ] shows the numerical estimates for temperatures at various tissue penetration depths obtained as a function of time @xcite .
this figure represents the hyperbolic ( solid line ) and parabolic ( dashed line ) temperature evolution at the four locations @xmath66 and @xmath67 mm for the thermokeratoplasty technique .
we can mainly find two differences between both solutions : firstly , according to the parabolic model , for the values @xmath68 and @xmath69 mm the maximum temperature has already been reached and is steadily decreasing . for @xmath70 and @xmath67 mm ,
the maximum temperature has not yet been reached and is initially increasing .
for the _ hhte _ predictions , we observe that for all penetration depths in some part of the plot ( or the entire plot ) the temperature is increasing .
the great differences in temperatures between both models ( overall at positions closer to the surface @xmath68 and @xmath69 mm ) is even more important when the energy application is pulsed , since in every new pulse the difference between the initial temperature could become greater and greater .
secondly , we can confirm that the behavior of both solutions is very different . at values @xmath68 and @xmath69 mm
, we notice a delay in the abrupt temperature drop associated with switching off the laser beam .
this delay is related to the finite thermal propagation speed in the _ hhte _ model . using in eq .
( [ speed ] ) the numerical value for diffusivity ( _ viz .
[ corneaparam ] ) and for the relaxation time @xmath71 predicts @xmath72 for the speed of these steps , which exactly agrees with the known finite speed of the thermal wave in the cornea @xcite .
this fact demonstrates the wave character of the _ hhte _ model . for depths
@xmath70 and @xmath67 mm , however , these drops are not observed , because they occur outside the range of the time interval considered . radiofrequency heating ( _ rfh _ ) is a surgical procedure broadly employed in many clinical areas such as the elimination of cardiac arrhythmias , the destruction of tumors , the treatment of gastroesophageal reflux disease and the heating of the cornea for refractive surgery @xcite . in the remainder of this section
we will focus on a model of the corneal _ rfh _ treatment .
the schematic diagram of the model geometry is shown in fig . [ modelpic ] . here
, a spherical electrode of radius @xmath73 is completely embedded and in close contact with the biological tissue , which has infinite dimension . in this case , we can also use a one - dimensional model ( @xmath74 being the spatial variable ) , since the geometry under consideration displays radial symmetry . again , assuming @xmath38 in the bioheat equation ( [ bioheat ] ) , the governing equation results from the combination of eq .
( [ hhte ] ) and the explicit form of the ( surgical ) heat source @xmath75 which in the case of _ rfh _ is a product of a radial @xcite and a temporal part @xmath76 so that the total applied power outside of the electrode is @xmath77 where @xmath78 is the usual solid angle subtended by a surface .
thus , the power ( flux ) per unit solid angle , @xmath79 , satisfies the usual @xmath80 intensity law .
note that @xmath46 is included to model a non - pulsed source .
as a result , for the corneal _ rfh _ the governing equation and the boundary conditions are : @xmath81\ ] ] @xmath82 \lim\limits_{r\to\infty } t(r , t)=t_0 & & \qquad\forall t>0 \\[.25 cm ] \displaystyle { { \partial^2t\over\partial t^2}(r_0,t ) + { 1\over\tau}\,{\partial t\over\partial t}(r_0,t ) = { 3k\over\rho_0 c_0 r_0\tau}\,{\partial t\over\partial r}(r_0,t ) } & & \qquad\forall t>0 \label{electroflux } \end{array}\ ] ] the last condition in eqs .
( [ electroflux ] ) contains the density and specific heat of the electrode ( where @xmath83 ) , given by @xmath84 and @xmath85 , respectively .
this condition is based on the fact that in the interior and on the surface of the spherical electrode it must hold @xmath86 which just states that the heat flux @xmath87 spreading into the biological tissue is due to the change of the thermal energy content of the electrode . applying gauss law to eq .
( [ eleccons ] ) and integrating over the entire volume @xmath83 yields @xmath88 where we have assumed that the electrode has relatively small radius and sufficiently high thermal conductivity , which makes the electrode act like a punctual heat source . in the derivation of eq .
( [ rfhcond ] ) , this justifies taking @xmath89 to be constant and pulling @xmath90 out of the volume integral . finally , by substituting the expression for the hyperbolic heat flux eq .
( [ hypflux ] ) into eq .
( [ rfhcond ] ) and using @xmath91 , where @xmath92 is the usual radial unit vector , one readily obtains the last condition for the electrode in eqs .
( [ electroflux ] ) .
the analytical solution of this problem and the corresponding _ phte _ solution have been obtained in ref .
@xcite as an application for corneal _
rfh_. we have visualized the data generated from ref .
@xcite in fig .
[ rfhpic ] in order to discuss the main characteristics of the model .
figure [ rfhpic ] represents the hyperbolic ( solid lines ) and parabolic ( dashed lines ) temperature profiles along the radial axis for the times @xmath93 and @xmath94 ms .
similar to the laser - heating case , for small times and locations near the electrode surface the temperatures from the _ hhte _ are greater than from the _ phte_. this fact also becomes important when instead of a non - pulsed _ rfh _
application a pulsed power is applied .
moreover , the presence of pronounced peaks in fig . [ rfhpic ] reveals the wave character of the _ hhte _ model and its prediction of a finite heat conduction speed @xmath95 .
in this paper , we have outlined the fundamentals and main differences between the parabolic and hyperbolic model of heat conduction .
these differences encourage the use of the _ hhte _ approach in processes in which great amounts of heat are transferred to any material in very short times .
specifically , laser and radiofrequency heating are two surgical techniques of this type of processes . through the application of _ hhte _ to these surgical techniques
, we have shown the main characteristics of this model and its differences with _
predictions . at the moment
, we are working on the analytical modeling of the hyperbolic bioheat equation including a source term for blood perfusion ( _ viz . _ eq .
( [ bioheat ] ) )
. this will be less important for non - perfused tissues ( such as the cornea ) , but relevant for well - perfused organs such as _ e.g. _ the liver .
on the other hand , we are studying the implications of the hyperbolic heat equation for the case of pulsed _ rf _ applications .
energy pulses for surgical procedures are being employed in such different areas as conductive keratoplasty ( _ ck _ ) and _ rfh _ to destroy tumors . however , we are well aware of the restrictions of purely analytical models , mainly due to the lack of mathematical tractability . for this reason
, we are planning to develop alternative theoretical models based on numerical techniques ( such as finite elements ) in order to use the hyperbolic bioheat equation in models with more complicated geometry ( more realistic electrode and tissue geometries with irregular boundaries ) and also varying thermal characteristics .
f. manns , d. borja , j.m .
parel , w. smiddy and w. culbertson , semianalytical thermal model for subablative laser heating of homogeneous nonperfused biological tissue : application to laser thermokeratoplasty , _ journal of biomedical optics _ * 8*(2 ) , 288297 , ( 2003 ) .
m. trujillo , m.j .
rivera , j.a .
lpez molina and e.j .
berjano , analytical thermal - optic model for laser heating of biological tissue using the hyperbolic heat transfer equation , submitted to _ mathematical medicine and biology _ , ( 2008 ) .
berjano , j.l .
ali and j. saiz , modeling for radio - frequency conductive keratoplasty : implications for the maximum temperature reached in the cornea , _ physiological measurement _ * 26*(3 ) , 15772 , ( 2005 ) .
a. erez and a. shitzer , controlled destruction and temperature distributions in biological tissues subjected to monoactive electrocoagulation , _ journal of biomechanical engineering _ * 102 * , 4249 , ( 1980 ) .
lpez molina , m.j .
rivera , m. trujillo and e.j .
berjano , effect of the thermal wave in radiofrequency ablation modeling : an analytical study , _ physics in medicine and biology _ * 53*(5 ) , 14471462 , ( 2008 ) .
lpez molina , m.j .
rivera , m. trujillo , f. burdo , j.l .
lequerica , f. hornero and e.j .
berjano , assessment of hyperbolic heat transfer equation in theoretical modeling for radiofrequency heating techniques , _ the open biomedical engineering journal _ * 2 * , 2227 , ( 2008 ) . in a physical system ( @xmath96 simply - connected domain ) is related to both , temperature change and divergence of the heat flux @xmath87 .
the coefficient @xmath97 gives the volumetric heat capacity in terms of thermal conductivity and diffusivity . ] and
cross - sectional surface area @xmath98 .
the corresponding flux , thermal energy per time and area , is proportional to the temperature gradient @xmath99 .
the sign indicates that heat flows from @xmath100 to @xmath101 . ] and
ambient temperature @xmath103 .
the laser pulse has a duration of @xmath104 with an incident energy flux @xmath105 . the represented curves for the _ hhte _ ( solid lines ) and fourier ( dashed lines ) model correspond to penetration depths @xmath106 .
the relaxation time is @xmath71 . , width=491 ] , and @xmath94 ms of a non - pulsed _ rf _
application with total power @xmath108 and duration @xmath109 .
the initial temperature is @xmath110 and the electrode radius @xmath111 .
the relaxation time is taken @xmath112 .
, width=491 ] | in modern surgery , a multitude of minimally intrusive operational techniques are used which are based on the punctual heating of target zones of human tissue via laser or radio - frequency currents . traditionally , these processes are modeled by the bioheat equation introduced by pennes , who considers fourier s theory of heat conduction .
we present an alternative and more realistic model established by the hyperbolic equation of heat transfer . to demonstrate some features and advantages of our proposed method
, we apply the obtained results to different types of tissue heating with high energy fluxes , in particular radiofrequency heating and pulsed laser treatment of the cornea to correct refractive errors . hopefully , the results of our approach help to refine surgical interventions in this novel field of medical treatment . , , , , and heat models , fourier heat equation , parabolic heat equation , bioheat equation , radiofrequency surgery , laser ablation |
sonoluminescence ( sl ) is an intriguing phenomenon , which consists of light emission by small collapsing bubbles inside liquids @xcite .
such bubbles are created by ultrasonic waves when the pressure of the liquid is reduced relatively to the pressure of the gas present in the medium @xcite . in general , the appearance of a cavity within a liquid implies the existence of a surface ( the wall ) dividing the region into two parts , each one occupied by a fluid : the inner cavity consisting of a gas and/or liquid vapor , and the liquid portion outside .
the bubble initially expand to a maximum volume , and then collapse . at the final stages of the collapse ,
the gas inside the bubble radiate light .
it is observed that the light emission is enhanced when atoms of noble gas are present inside the bubble . in the 90s , gaitan et al .
@xcite obtained in laboratory the necessary conditions to create and trap a single bubble the size of a few microns , levitating in a bottle of water under the action of a strong , stationary - wave sound field , emitting periodically flashes of light in each acoustic cycle .
the trapping of a single sonoluminescent bubble in the liquid , yielding what is called single - bubble sonoluminescence ( sbsl ) , and the production of repeated cycles of expansion and contraction , excited by ultrasonic acoustic waves , allowed a more accurate study of the phenomenon .
this stability of the bubble made possible more detailed studies about the duration of the flash of light and the size of the bubble .
the conversion of the energy of sound waves into flashes of light occurring in such bubbles with few microns in size is an interesting field of research , if one considers the difficulties inherent to the process , the inadequacy of some theories and models proposed , and experimental limitations . despite the existence of a wide variety of theories and models , the phenomenon is not completely explained , and there are still many open questions . among them , we can cite the heating mechanism of the gas inside the bubble ; the process of light emission ; the role of temperature of the liquid in the intensity of the emitted radiation ; the reason why water is one of the ideal fluids for observing the phenomenon .
up to now , the phenomenon has been extensively studied , but its detailed mechanisms remain still unclear .
the study of sl involves topics of physics such as hydrodynamics and thermodynamics , besides electromagnetism , statistical mechanics and atomic physics if one wants to go into the emitting mechanism .
the dynamics of the bubble can be described by the navier - stokes equation @xcite , provided the initial state of the bubble is defined by a set of physical parameters .
the energy of one photon emitted in sbsl , if compared with the energy of one atom vibrating in the sound wave which gave rise to it , typically provides a value of of order @xmath0 , which shows the high focalization of energy in this effect @xcite .
the temperature in the interior of the bubble can reache thousands of degrees kelvin during the collapse phase @xcite . in this work
we carry out a modeling of sbsl hydrodynamics , using the radius of the bubble as the variable of interest .
there are several models that describe the time evolution of the radius of the bubble .
the most common are : the rayleigh - plesset model , the herring - trilling model , and the keller - miksis model @xcite .
all of these models are derived from the navier - stokes equation , using different simplifying assumptions @xcite .
rayleigh - plesset equation describes the behaviour of compressible or incompressible fluids , and can therefore be used to compare the effects of the compressibility of the liquid on the time evolution of the bubble radius .
other physical parameters like the liquid viscosity , the properties of the gas inside the bubble , and the superficial tension of the wall can be taken into account in these models . in order to compare the effects of these factors ,
we define in this work an useful parameter , the _ damping factor _ , which is defined as the ratio between the first and second highest values of the bubble radius .
this parameter is relevant , since it compares two values of the radius : one before and the other one after the light emission .
a detailed study of the bubble behaviour , including the variation of the wall speed as a function of time , is also carried out .
this work is addressed mainly to ( under)graduate students and teachers .
the subject requires the domain of calculus and many concepts from various fields of physics at intermediate level
. the work can be useful in physics courses at university level , as well as an illustrative example of numerical calculus applied to a highly nonlinear phenomenon .
it may also be valuable as an introductory - level text on sbsl addressed to young researchers .
the work is the result of a research project in sonoluminescence carried out together with undergraduate students .
the behaviour of a nonrelativistic fluid is described by the navier - stokes equation , which is valid at each point of the fluid and can be written as @xmath1 where @xmath2 is the fluid mass density , @xmath3 is the velocity field inside the fluid , @xmath4 is the resultant of body forces ( e.g. , gravity ) per unit volume of the fluid , @xmath5 is the pressure field inside the fluid and @xmath6 is the fluid viscosity .
the term @xmath7 on the left - hand side of the above equation is the material derivative of the velocity of the fluid element .
the material derivative is given by the operator @xmath8 where the first term on the right - hand side is the time derivative with respect to a fixed reference point of space ( the euler derivative ) , and the second term represents the changes of the field velocity along the movement of the fluid .
the quantity @xmath9 on the right - hand side stands for the acceleration originated from the body forces acting on the fluid element , such as gravity or electromagnetic forces , for example .
the second and third terms represent respectively the hydrostatic force and the viscous force , both per unit volume .
the form navier - stokes equation is written above means that we are considering the fluid as newtonian .
now , we remember that the flow inside and around the bubble is restricted to the radial direction , in other words , the problem exhibits spherical symmetry , which in fact is valid even beyond the neighborhood of the bubble , provided the shape of the flask is spherical , a typical experimental situation .
thus , we take into account only the expansion and contraction motion of the bubble s radius . besides , we assume as a first approach that the compressibility of the liquid is much smaller than that of the gas inside the bubble . in this case , one derives from eq .
( [ 1 ] ) the rayleigh - plesset equation ( a dot stands for one time derivative , @xmath10 , etc ) : @xmath11 where @xmath12 is the bubble s radius , @xmath13 is the variable gas pressure inside the bubble ( @xmath13 is assumed to be uniform in our model ) , @xmath14 is the pressure of the liquid measured at any remote point from the bubble ( typically , @xmath15 atm ) , @xmath16 is the driven acoustic pressure at the point where the bubble is placed , and @xmath17 is the liquid surface tension at the bubble wall .
@xmath16 is assumed to be sinusoidal and starting an expansion cycle in @xmath18 , that is , @xmath19 @xmath20 being the amplitude of the driven pressure and @xmath21 the ultrasound angular frequency in resonance with the natural oscillations of the flask , such that the driven pressure generates a stationary ultrasound wave that traps the bubble at its center , on a pressure antinode .
often , the effects of the compressibility of a liquid can be neglected in many problems of hydrodynamics . however , in the case of sbsl this approach is no longer justified , because a large amount of the acoustic energy driven to the bubble is emitted back from it to the liquid , in the form of a spherical shock wave ( only a small amount is in fact converted into light ! ) , which obviously could not exist in an incompressible medium ( the acoustical wave emitted by the bubble is experimentally important , since its detection by a hydrophone signals the presence of the trapped bubble at the center of the spherical flask ) .
it follows that when the compressibility of the liquid is considered , a new term is added to the right - hand side of eq .
( [ 3 ] ) , leading to the modified rayleigh - plesset equation @xcite @xmath22 where @xmath23 is the speed of sound in the liquid ( which henceforth we will assume to be water ) .
we adopt a van der walls equation of state for describing the gas pressure inside the bubble , which reads @xmath24 where @xmath25 is the static bubble radius , that is , the ambient bubble radius when it is not acoustically forced , @xmath26 is the characteristic van der waals hard - core radius of the gas inside the bubble and @xmath27 is the ratio between the specific heats of the gas at constant pressure and at constant volume ( the adiabatic index ) .
so , the gas pressure varies with time only by means of the bubble radius , @xmath12 .
it was assumed in eq .
( [ 3.2 ] ) that the gas undergoes a so fast cycle of expansion and collapse that it is adiabatic
. however , a more accurate analysis @xcite allows one to conclude that the expansion is approximately isotermic ( @xmath28 ) and only the final part of the collapse is indeed adiabatic . our simplified model here consider the whole cycle as adiabatic . in eq .
( [ 4 ] ) , the time derivative of the gas pressure is explicitly given by @xmath29 equations ( [ 3 ] ) and ( [ 4 ] ) are second - order differential equations for the radius @xmath12 , which have no analytical solution .
so , in order to solve numerically these equations for a given set of system parameters , we used the mathematical program matlab ( the corresponding program codes can be requested by email to any of the authors ) .
let us first consider the case in which the compressibility of the liquid is neglected .
figure [ fig:100 ] shows the solution of eq .
( [ 3 ] ) for a given set of parameters describing some properties of the gas and the liquid , and at forcing pressure @xmath30 atm .
the parameters used are shown in table [ table1 ] , where the noble gas argon is considered because it is well - known in sbsl literature @xcite that , after some cycles , it is the only remaining constituent of the initial air bubble , due to chemical reactions that the other components participate , which carry them to the water outside . the solution shown in fig .
[ fig:100 ] is qualitatively similar to the experimental results @xcite , but quantitatively very different : the experimental afterbounces are much smaller than those that appear in fig . [ fig:100 ] .
this is in agreement with the fact that indeed a considerable amount of energy is lost in the first contraction ( collapse ) , which , as we have already pointed out , is due to sound radiation by the bubble , an effect that is impossible theoretically under the assumptions that led to eq .
( [ 3 ] ) .
.[table1 ] parameters used in the numerical simulation .
[ cols= " < , < " , ] we can see from table [ table2 ] that the superficial tension strongly affects the dynamics of the bubble , when compared with the other two properties considered here , since it substantially reduces the damping factor . according to the experimental results existing in the literature for sbsl @xcite , the damping factor lies between the two extreme values shown in the table , and closer to the lower values , thus pointing out the importance of including the surface tension in the calculations .
in this work we have presented a model describing a single bubble sonoluminescence .
we obtained solutions for the equations of the model in the cases of compressible and incompressible liquids . comparing figs .
[ fig:100 ] and [ fig:200 ] with the experimental results , we show that the modified rayleigh - plesset equation valid for a compressible liquid , given by eq .
( [ 4 ] ) , describes the bubble dynamics with a better accuracy @xcite , which is due to the experimentally observed acoustic radiation from the bubble , a fact that demands the liquid compressibility to be different from zero .
we introduced the damping factor in order to compare the effects of some properties of the system on the dissipation of energy of the bubble after the collapse ( first compression ) .
we pointed out that the superficial tension of the liquid can not be neglected in any model describing sbsl dynamics , as we summarized in table [ table2 ] .
it is worth to mention that chemical effects was not considered in the present work .
the chemical factors clearly affect the thickness of the bubble wall , thereby also contributing to the generation of a shock wave .
of course , a more realistic treatment must include such kind of ingredient .
also , as mentioned above , the process that the bubble gas undergoes is not adiabatic during the whole cycle , as we considered here , but instead isotermic during the expansion , which affects the value of the exponent @xmath27 . however , as pointed out in ref .
@xcite , just modifying the numerical calculation in order to interpolate between @xmath31 and @xmath32 is not sufficient to completely describe the behaviour of @xmath12 in all of its details .
matlab routines that generates our results can be requested by email to any of the authors and we hope that the simulations presented here will be valuable for teachers , ( under)graduate students and young researchers in order to have a first contact with single - bubble sonoluminescence and some of its remarkable features .
the authors thank the brazilian foundations cnpq , faperj and capes for the financial support . | sonoluminescence ( sl ) is the phenomenon in which acoustic energy is ( partially ) transformed into light .
it may occur by means of many or just one bubble of gas inside a liquid medium , giving rise to the terms multi - bubble- and single - bubble sonoluminescence ( mbsl and sbsl ) . in the last years
some models have been proposed to explain this phenomenon , but there is still no complete theory for the light emission mechanism ( especially in the case of sbsl ) . in this work , we will not address this more complicated particular issue , but only present a simple model describing the dynamical behaviour of the sonoluminescent bubble , in the sbsl case .
using simple numerical techniques within the software matlab , we discuss solutions considering various possibilities for some of the parameters involved : liquid compressibility , superficial tension , viscosity , and type of gas
. the model may be used as an introductory study of sonoluminescence in physics courses at undergraduate or graduate levels , as well as a quite clarifying example of a physical system exhibiting large nonlinearity . |
although the physical origins of a number of solar radio burst types are well known , e.g. , type ii and type iii bursts ( krger @xcite ; bastian et al .
@xcite ) , there are still many questions which types of the radio plasma emission emerge directly from the primary energy release of solar flares , which is supposed to result from magnetic reconnection in coronal current sheets ( e.g. , priest @xcite ; masuda et al.@xcite ; tsuneta @xcite ) . furthermore , there are still different viewpoints regarding the question in which frequency range this radio emission can be observed .
decimetric continuum radio bursts are often related to the impulsive phase of solar flares , and their frequencies correspond to the densities supposed to exist in the primary energy release volume .
these bursts are usually regarded to result from a distribution of flare - accelerated particles trapped in a coronal magnetic loop ( kuijpers @xcite ) . in the trap ,
the particles develop a loss - cone distribution function , which is unstable and supports a plasma emission process that leads to fundamental emission near the electron plasma frequency , @xmath0 , or to harmonic emission near @xmath1 . according to this model ,
the bursts carry signatures not only of the particle acceleration in the primary energy release process , but also of the evolution of the trapped particle distribution in velocity space and of the evolution of the loop during the flare .
continuum radio bursts , especially those in the decimetric wavelength range , often exhibit broad - band pulsations of the flux .
the timescales lie typically in the range @xmath2 s. proposed mechanisms of the pulsations fall into three categories : ( 1 ) mhd oscillations of a magnetic loop , which modulate the radio emissivity , ( 2 ) an oscillatory nonlinear regime of the kinetic plasma instabilities that emit the radio waves , and ( 3 ) a modulation of the acceleration of particles which represent the source of the wave energy ( aschwanden @xcite ) .
the third of these includes non - stationary magnetic reconnection .
improved knowledge of magnetic reconnection in current sheets is imperative for resolving the flare problems .
such studies can be guided by radio observations , which often carry details of the dynamical plasma processes not visible at other wavelengths . in numerical simulations of reconnection ( e.g. , ugai & tsuda @xcite ; sato & hayashi @xcite ; karlick @xcite ; scholer @xcite ; ugai @xcite ) , as well as in the interpretation of soft x - ray observations ( tsuneta @xcite ) , the focus has usually been on stationary petschek - type reconnection . however , temporal fine structure is abundant in radio bursts and often also present in hard x - ray bursts ( e.g. , kiplinger et al .
another problem lies in the fact that numerical simulations have to be performed in rather small boxes and over rather short time intervals .
there is a permanent need to extend the simulations further in space ( where they tend to reveal stronger dynamics ) and in time , and attention must be devoted to the re - scaling of the numerical results when specific observations are to be interpreted . in the present paper
we consider a pulsating decimetric continuum radio burst event .
its peak was coincident with the impulsive phase of a solar flare ( october 5 , 1992 ) .
rather detailed knowledge of this flare is already available from a thorough analysis of its soft x - ray emission , which includes , in particular , the ejection of a plasmoid ( ohyama & shibata @xcite ) .
we argue that the pulsations can be explained by modulations of particle acceleration in a highly dynamic regime of magnetic reconnection in an extended current sheet above the soft x - ray flare loop and that the radio emission process does not necessarily require trapping in a magnetic loop .
the time profile of the radio burst appears to reflect the dynamics of the reconnection process directly .
mhd simulations of reconnection in a long current sheet are presented to support the model .
these simulations provide also further insight into the process of plasmoid formation in a current sheet , which is directly related to episodes of enhanced reconnection and enhanced particle acceleration .
the observations are presented in sect .
[ obs ] and our mhd simulations are presented and scaled to coronal conditions in sect .
[ reconn ] .
a discussion of the new interpretation of pulsating continuum bursts and the conclusions are given in sects . [ discuss ] and [ concl ] , respectively .
the flare on october 5 , 1992 , 9:25 ut was an impulsive , compact - loop , m2 class event that took place at the west limb , probably in active region noaa 7293 , which was located slightly behind the limb .
its hard x - ray lightcurve and the decimetric dynamic radio spectrum are plotted in fig .
[ main ] along with two characteristic single - frequency decimetric flux profiles .
[ microwave ] shows the corresponding microwave burst .
the hard x - ray ( hxr ) emission showed a complex temporal evolution .
the flare started with an extended period of low - level emission .
the first impulsive rise near 9:24:25 ut was followed by several strong enhancements , each of @xmath3 s duration ( fwhm ) .
the spectrum was hardest near 9:25:05 ut . the yohkoh / hxt 1423kev and 2333kev band images ( 9:24:289:25:16 ut ) show a single , slightly elongated source with centroid position just above the west limb ( sato et al .
the weak microwave burst had a time profile similar to the hxr burst , but was much smoother .
also the goes soft x - ray lightcurves ( solar geophysical data ) show an impulsive rise near 9:24:25 ut , following a period of low - level activity .
the decimetric radio burst consisted of continuum emission superimposed with strong broadband pulsations during its whole lifetime . the general flux level varied in a complex manner during the event , showing several enhancements , as is typical of decimetric bursts .
the burst as a whole drifted towards lower frequencies at a rate @xmath4 mhzs@xmath5 .
the main burst phase started at 9:24:27 ut at a frequency slightly higher than the preceding pulsations and with a temporarily much higher drift rate of @xmath6 mhzs@xmath5 .
the average drift rate decreased toward the end of the event .
the burst was weakly polarized , which is typical of type iv bursts from sources near the solar limb .
it is not known whether the emission was fundamental or harmonic radiation .
no other bursts were recorded at this time at frequencies below 0.6 ghz ( tremsdorf and zrich radiospectrographs ) .
[ pulsedetail ] displays the pulsations during the early phase of the burst in greater detail .
the intervals between the pulses and their individual durations were not regular , leading to a quasi - periodic appearance of the flux variations .
the timescales of the pulses were lying in the range 0.53 s , and the fwhm of the auto - correlation was 1.3 s. two additional emission bands of low intensity existed during part of the burst ( fig .
[ threebands ] ) .
the high - frequency band appeared only during a short interval ( 09:24:5525:05 ut ) .
its drift and intensity variations were not correlated with with the main band .
the low - frequency band , which had an average frequency drift similar to the main band and weakly correlated intensity variations , could be discerned during @xmath7:24:2009:25:25 ut .
it is not clear whether it then decayed , or merged with the main band , or continued to drift but was masked by the low - frequency edge of the main band , which drifted to @xmath8 ghz at 09:26:18 ut .
the hxr lightcurve does not show much structure at short timescales ( most of the fluctuations seen at the resolution of 64 ms in fig .
[ main ] are noise ) , but the auto - correlation ( e.g. in the regular interval 09:23:28 to 09:24:13 ut ) shows a fwhm of 4.5s , indicating significant substructure in time .
the 2550kev batse lightcurve of the event considered here had been included in an earlier search for significant timescales in hard x - ray time profiles using a wavelet analysis method ( aschwanden et al .
@xcite ) , and a weak contribution around 4 s had been found in addition to the dominant timescale of @xmath9 s , which reflects the overlapping main pulses seen in fig .
[ main ] .
hxr lightcurves are generally composed of a direct - precipitating component , which carries the information on the rapid variations of the acceleration process , and a smoothed component , leaking out of a trap , that has lost this information . a trap model could indeed be fitted to energy dependent time delays in the low - pass filtered hxr lightcurves of the considered flare ( aschwanden et al .
this showed the importance of trapping in this flare and suggested a density in the trap of the hxr - emitting particles of order @xmath10 @xmath11 , which is roughly in accordance with the density of the flare loop estimated from the soft x - ray data ( see below ) .
there was a high correlation between the _ general evolution _ of the decimetric burst and the hxr lightcurve during the initial phase : the period of low - level emission and the impulsive onset of the main phase were coincident .
the correlation became much weaker as the decimetric flux decreased in the course of the event .
the strongest and hardest hxr peak was accompanied by only rather weak decimetric emission , but one of the main bunches of enhanced decimetric pulses later on ( @xmath12:25:50 ut ) was again nearly simultaneous with a significant hxr enhancement .
the correlation of the general evolution shows that the decimetric burst was closely associated with the first impulsive energy release of the flare and with the processes that led to that energy release .
differences to the hxr burst in the later phases do not necessarily imply a loss of association , since the radio emission in general reflects not only the production of nonthermal particles but is also sensitive to changes in the magnetic field , which influence the slope of the particle distribution function and , hence , the growth rate of the plasma waves , and to density changes in the source and along the ray path , which influence the propagation of the radio waves .
it is , however , quite probable that the strongest and hardest hxr peak resulted from flaring in a part of the active region that was not magnetically well connected with the decimetric source .
the situation here is typical of comparisons between decimetric and hxr time profiles , where an association can often be established only from the coincidence of the main features like onset , main rise , or peak ( aschwanden et al .
@xcite ) .
a _ detailed comparison _ by cross - correlation was made for the initial , relatively regular part ( shown in fig .
[ pulsedetail ] ) .
the time resolution of the batse data in this pre - trigger interval is 1.024s , so the correlation refers to structures of at least 2s duration , which in general do not correspond to single decimetric pulses but to small groups .
the cross - correlation coefficient has a distinct minimum of @xmath13 near zero lag , indicating a negative correlation ( anti - correlation ) .
its fwhp of only 3.0s suggests a trend for individual hxr pulses to peak when the radio flux reaches a minimum .
the same property appears also in the main part of the event after subtracting a gliding - average background with 10s or 20s time constant .
the positive correlation of the general evolution of the decimetric and hxr lightcurves indicates the acceleration of the radiating electrons by the same global destabilization of the magnetic field configuration , and the anti - correlation of the details suggests acceleration by the same or by closely related spatial substructures with a possibly complex relationship between the radio and hxr sources .
the soft x - ray images of the event , analysed by ohyama & shibata ( @xcite ) , revealed a nearly stationary low - lying flare loop and a plasmoid ejection , which was first seen at 9:24:40 ut at a height of @xmath14 km .
in addition , a large , faint loop was apparently connected with the rising plasmoid .
the plasmoid was continuously accelerated and reached velocities of @xmath15 kms@xmath5 at 9:25 ut and @xmath16 kms@xmath5 at 9:26 ut . temperatures and plasma densities in the soft x - ray structures were estimated by ohyama & shibata ( @xcite ) as follows : ( a ) in the flare loop at 9:24:189:24:50 ut : @xmath17 mk and @xmath18 @xmath11 and at 9:24:589:25:58 ut : @xmath19 mk and @xmath20 @xmath11 ; ( b ) in the plasmoid at 9:25:00 ut : @xmath21 mk and @xmath22 @xmath11 ; and ( c ) in the expanding loop at 9:26:08 ut : @xmath23 @xmath11 under the assumption of @xmath24 mk and @xmath25 @xmath11 for @xmath26 in the range 414 mk .
parameters in the current sheet , supposed to exist between the nearly stationary flare loop and the rising plasmoid , were not explicitly given but can be estimated from fig .
7 of ohyama & shibata ( @xcite ) as @xmath27 @xmath11 and @xmath28 k at 9:25:18 ut .
( the latter values may actually refer to the rear side of the plasmoid , which formed an elongated structure . ) in comparison , the most intense radio emission in the 1.11.6 ghz frequency range implies source densities in the range @xmath29 @xmath11 in case of fundamental or @xmath30 @xmath11 in case of harmonic radiation .
although the uncertainties are substantial , it is apparent that the densities in the radio source correspond more closely to the densities in the plasmoid and in the current sheet than to the densities in the flare loop or in the expanding loop ( for the reasonable assumption of @xmath31 mk ) .
the substantial increase of the density in the flare loop during the event rules this loop out as site of the radio source , since the radio emission showed , on average , a negative frequency drift .
interpretations of the pulsations in the considered event in terms of existing models for the modulation of a magnetic loop by mhd waves can be ruled out , because these models predict pulsations with regular or slowly varying period and most of them also with constant or exponentially decaying amplitude ( see aschwanden @xcite ) , while the observed intervals between the pulses , the durations , and the peak intensities were all varying rapidly and partly irregularly .
the geometry of this _ eruptive _ flare and the observed relation between hxr and decimetric radio pulsations are not suggestive of modulations by oscillatory nonlinear instabilities of electrons stably trapped for more than a minute .
it appears more obvious to relate the pulsations to changes of the plasma environment and to the corresponding modulations of the particle acceleration or radio emissivity than to oscillatory nonlinear kinetic processes in an idealized stationary trap or current sheet model .
the yohkoh observations provide strong evidence that a current sheet lying above the soft x - ray flare loop and an outward propagating plasmoid were essential elements of the flare under consideration .
the basic magnetic topology appears to be similar to the 2d kopp - pneuman model of flares ( see , e.g. , tsuneta @xcite ) .
estimates of the energy content of the plasmoid suggested that it did not contain sufficient energy to form the current sheet by sweeping up an overlying closed magnetic field structure ( ohyama & shibata @xcite ) .
thus , either the plasmoid was formed and accelerated by processes operating in the preexisting current sheet , or it was formed jointly with the current sheet by a global mhd instability . considering the limited frequency extent and average frequency drift of the observed radio burst , and the estimated densities , we interpret the burst as emission originating from the plasmoid ( or from dense plasmoid - like structures in the current sheet ) .
the frequency drift of the burst can then be attributed to the ejection of the plasmoid ( or to the corresponding upward expansion of the current sheet ) . to explain the radio pulsations , we focus on small - scale mhd processes within the current sheet , disregarding aspects of interaction with the dense flare loop lying below the current sheet . we expect that the small - scale dynamics in the current sheet will be similar in the two variants of plasmoid formation , which were mentioned above , since the magnetic topology and the large - scale plasma flows become similar , once a plasmoid is formed . here
we will concentrate on the former case , i.e. , spontaneous reconnection triggered by an initial perturbation within the current sheet ; no driving flows are prescribed at the boundaries of the numerical system .
we note that such flows could not be specified from the existing observations of the event .
a two - dimensional ( 2d ) model appears to be well justified to study effects intrinsic to the current sheet , since the size of the plasmoid along the line of sight was estimated to be of order @xmath32 km , which exceeds the supposed current sheet width by several orders of magnitude .
although the expanding loop - shaped structure , if it was really connected with the plasmoid , implies the presence of a magnetic guide field component , we restrict ourselves here further to the case of vanishing guide field , i.e. , an initially antiparallel ( `` neutral '' ) current sheet is studied .
as long as the internal current sheet dynamics including the formation and acceleration of the plasmoid are concerned , this simplification also appears reasonable within the 2d approximation for the following reason . in 2d incompressible mhd ,
an initially homogeneous guide field is only passively advected , it has no influence on the dynamics . in 2d simulations of compressible current sheet dynamics ,
its influence has been found to remain very small ( ugai @xcite ) .
( a guide field becomes essential in 3d reconnection studies , where it can lead to the formation of structures like interlinked flux tubes , which are otherwise absent . )
magnetic reconnection is often considered in the framework of the stationary petschek model ( e.g. , ugai @xcite , @xcite ; yokoyama & shibata @xcite ; tsuneta @xcite ) and its generalizations ( priest & forbes @xcite ) .
however , there is a long - lasting dispute whether stationary petschek reconnection is possible ( biskamp @xcite ; forbes & priest @xcite ; scholer @xcite ; kulsrud @xcite ) . for long current sheets , numerical simulations indicated that resistive magnetic reconnection proceeds in a time - variable manner , involving repeated formation of magnetic islands and their subsequent coalescence ( biskamp @xcite ; scholer & roth @xcite ; karpen et al .
this is particularly true if the resistivity is permitted to vary .
the process is known as `` secondary tearing '' or as the `` impulsive bursty '' regime of reconnection ( priest @xcite ) .
the repeated coalescence of islands may lead to the formation of a growing , and eventually large - scale , plasmoid ( scholer & roth @xcite ; schumacher & kliem @xcite ; magara et al .
coincident with the formation of each new magnetic island , the electric field at the new magnetic x point is enhanced .
this probably leads to a burst of accelerated particles ( although the detailed mechanisms of particle acceleration in reconnection regions still remain unknown ) . in the simplest model
, each formation of a new magnetic x point corresponds to a burst of acceleration , which , in its turn , corresponds to an enhancement of the radio flux ( one pulsation ) .
more complicated schemes are conceivable and briefly discussed in sect .
[ discuss ] .
the equations of compressible , resistive one - fluid mhd used in this paper are @xmath33 where the current density * j * , the total energy density @xmath34 , and the flux vector @xmath35 are given by @xmath36 and @xmath37 is the internal energy per unit mass , which is related to the pressure through the equation of state , @xmath38 .
in the two - dimensional model adopted here , @xmath39 for all quantities and @xmath40 ; hence the current density and the electric field possess only @xmath41 components .
further , the ratio of specific heats is @xmath42 , where @xmath43 is the number of degrees of freedom .
the electric field is given by @xmath44 an antiparallel harris equilibrium ( a neutral current sheet ) with uniform density is chosen as the initial condition : @xmath45 where the plasma beta is defined as @xmath46 .
the variables are normalized by quantities derived from the current sheet half width @xmath47 and the asymptotic ( @xmath48 ) alfvn velocity @xmath49 of the configuration at @xmath50 .
time is measured in units of the alfvn time @xmath51 , and @xmath52 , * e * , * j * , and @xmath53 are normalized by @xmath54 , @xmath55 , @xmath56 , and @xmath57 , respectively
. then the lundquist number is given by @xmath58 .
the normalized variables will be used in the remaining part of this subsection .
the equilibrium is initially perturbed by a localized resistivity @xmath59\,,\ ] ] fixed for a few alfvn times , @xmath60 .
thereafter the resistivity is variable and is determined self - consistently at every grid point and time step from the electron - ion drift velocity @xmath61 ( where the factor @xmath62 results from the normalization and @xmath63 is the ion inertial length ) .
an `` anomalous '' value , @xmath64 , is set if a threshold , @xmath65 , is exceeded , otherwise vanishing resistivity is assumed : @xmath66 here @xmath67 and @xmath68 .
such a resistivity model is customary in mhd simulations of magnetic reconnection ( e.g. , ugai @xcite ; yokoyama & shibata @xcite ) .
we have verified that the dynamics of the system does not significantly depend on the particular choice of the parameters @xmath69 , @xmath70 , and @xmath65 in a substantial range around the values used here .
also the dependence on @xmath71 is weak .
simulations with @xmath72 show less variable secondary tearing .
we present a low - beta simulation , @xmath73 .
( [ eq_rho])-([eq_energy ] ) are transformed into a flux conserving form , @xmath74 , for each of the six integration variables , @xmath75 , where * f * and * g * are the nonlinear flux terms . a two - step lax - wendroff scheme according to ugai & tsuda ( @xcite ) and a uniform cartesian grid with resolutions
@xmath76 are employed for the numerical integration . in order to stabilize the scheme , artificial smoothing ( sato & hayashi @xcite )
is applied such that the variables at each grid point are replaced after each full time step of the lax - wendroff algorithm in the following manner @xmath77 with @xmath78 varying in the range 0.970.98 during the simulation . in order to study the effect of secondary tearing
, the numerical box must not be too small . for box lengths @xmath79 , which is a common value in the literature , secondary
tearing occurs only for rather small box heights , @xmath80 . increasing the box length to @xmath81 ,
leads to secondary tearing for all reasonable values of @xmath82 , however , the details of the time profile of the reconnection rate , in particular the development of irregular or quasi - periodic reconnection , depend on @xmath83 until @xmath83 exceeds a value @xmath84 .
the simulation presented here used @xmath85 and @xmath86 .
the simulation is carried out in the first quadrant , using symmetrical boundary conditions at the lower and left boundaries .
open boundary conditions are realized at \{@xmath87 } and \{@xmath88 } by the requirement that the normal derivatives of all variables vanish , except for the normal component of @xmath89 , which is determined from the solenoidal condition .
the dynamical evolution of the considered system , shown in figs .
[ bej_evol ] and [ upr_evol ] , starts as in many other investigations of spontaneous reconnection ( e.g. , ugai @xcite ) .
the initial perturbation causes a small burst of reconnection at the origin , which leads to acceleration of plasma along the @xmath90 axis by the lorentz force @xmath91 .
mass conservation then enforces an inflow @xmath92 into the region of the initial perturbation .
this convects new magnetic flux toward the sheet and causes the current density to increase again locally .
after @xmath93 , the threshold @xmath65 is exceeded and anomalous resistivity is spontaneously excited at the origin . this enables a positive feedback between reconnection , acceleration of the fluid , and the local rise of the resistivity . the continuing and initially amplifying reconnection leads to a strong outflow of heated plasma along the @xmath90 axis , which eventually reaches the alfvn velocity and drives a high - pressure region in front of it toward the outer boundary .
the state of the system during this phase of amplifying petschek - like reconnection is illustrated in figs .
[ bej_evol ] and [ upr_evol ] at @xmath94 and @xmath95 , where the current density ridges near the separatrix lines of the magnetic field delineate the slow mode shocks .
since the level of anomalous resistivity at the origin can adjust to changes in the outer region , it has often been assumed that a balance between convection and diffusion is reached and that , consequently , the petschek - like reconnection state , once formed , does approach stationarity ( e.g. , ugai @xcite ) . also many observational studies of solar eruptions have assumed that stationary petschek reconnection occurred ( e.g. , tsuneta @xcite ) .
however , stationary petschek reconnection has been obtained in numerical experiments only if a localized resistivity was held fixed in time or if the box length was small , @xmath96 ( see scholer @xcite ) .
variable resistivity leads to non - stationary reconnection if the current sheet is sufficiently extended , independent of whether the resistivity model is based on the electron - ion drift velocity or on the current density ( scholer & roth @xcite ) .
this has recently been confirmed for 3d reconnection as well ( schumacher et al.@xcite ) . the principal effect that leads to variable reconnection appears to be the inability of the fluid to carry a sufficient amount of magnetic flux into the strongly localized diffusion region to support the alfvnic outflow of reconnected flux in steady state ( kulsrud @xcite ) .
the free acceleration of the plasma , which is penetrated by reconnected field lines , into the @xmath90 direction causes the inflow to steepen the current density around the origin in a range of increasing extent .
although , according to eq .
( [ etaandef ] ) , the density depression at the origin tends to keep the anomalous resistivity more localized than the range of enhanced current density , also the area of anomalous resistivity excitation increases into the @xmath90 direction .
its extent reaches @xmath97 at @xmath98 . at the same time
, the width of the anomalous resistivity region in the @xmath99 direction stays approximately constant , @xmath100 .
the central part of the current sheet has thus formed a sweet - parker configuration , which implies a reduction of the reconnection rate in comparison to the petschek configuration ( @xmath101 ) .
the acceleration of the fluid is then also reduced locally , and both the current density peak and the density depression at the origin are reduced . a resistivity maximum remains at the edge of the elongated @xmath64 area , while the resistivity drops to zero between the new maximum and the origin .
this leads immediately to the formation of a new x point ( and its mirror image ) at @xmath102 the so - called secondary tearing ( scholer & roth @xcite ) .
the sequence then repeats at the new x point , where the level of the resistivity and the field line reconnection are initially also strongly amplified .
this drives the coalescence of the magnetic islands between the three x points , resulting in a plasmoid at the origin . over the next @xmath103 alfvn times ,
the magnetic structure at the new x point becomes flat and the resistivity area elongates and splits again , forming a new x point further out ( @xmath104 ) .
the corresponding magnetic island is again driven toward the central plasmoid by the rapidly evolving new x point . over a period of @xmath105 ,
seven strong new x points are created , supplemented by a number of weak x and o pairs which can be considered as `` fluctuations '' and are of little dynamical importance .
all seven magnetic islands merge with the central plasmoid , while some of the `` fluctuations '' merge with the outflow .
this merging is accompanied by transient excitations of anomalous resistivity between the approaching islands ( see fig .
[ bej_evol ] at @xmath106 ) . an integral view of the dynamics is given in fig .
[ r_rate ] , where the electric field at the various main x points ( i.e. , the reconnection rate ) and the location of the dominant x point in the system are plotted .
the seven new main x points are separated using different linestyles .
their individual variations contain imprints of the additional formations of weak x points , which disturb the magnetic configuration and the outflow from the main x points .
the `` dominant '' x point in the bottom panel is the one where instantaneously the highest reconnection rate occurs .
the change - over of maximum reconnection to a stronger new x point further out marks jumps in the position of the dominant x point .
it is apparent from the figure that the secondary tearing leads to higher reconnection rates than the petschek - like reconnection , which exists only during the first @xmath107 .
the reconnection rate in the numerical experiment is of the same order as that implied by the observations , @xmath108 ( ohyama & shibata @xcite ) .
the intervals between the reconnection peaks are not uniform , but also not completely irregular . as a result of secondary tearing
, the central plasmoid is continuously fed with heated material carrying reconnected magnetic flux .
the highly variable inflow into the plasmoid from the neighbouring x points triggers oscillations of the plasmoid as a whole .
these oscillations in their turn enhance the time variability of the highly nonlinear , threshold - depending reconnection process in the outer parts of the extended current sheet .
the pressure and density in the plasmoid are continuously rising , causing it to expand in the @xmath90 and @xmath99 directions .
this expansion appears to be the reason of the general slow decline of the activity seen in fig .
[ r_rate ] after @xmath104 .
the increasing width of the plasmoid in the @xmath99 direction makes the convection of new magnetic flux toward the neighbouring x point more difficult because an increasing amount of field line bending is required , and the space available to form an extended flat current sheet section between the plasmoid and the boundary @xmath83 is reduced .
this decrease of the average reconnection rate is expected to be far smaller in the solar corona , where the magnetic islands are much more elongated in the @xmath90 direction and less field line bending is required ( see sect .
[ scales ] ) .
we have seen that the dynamical evolution leads to two types of dense structures .
the first one is the symmetric pair of high - pressure regions driven ahead of the outflow , which commences with the initial perturbation .
this structure is commonly referred to as plasmoid , although only the field lines at its rear side are closed , while the field lines at its front side close at infinity .
( in the numerical experiment , a small fraction of the flux closes also at its front side , due to the nonvanishing numerical diffusion . )
the second one is the plasmoid resulting from secondary tearing .
[ plasmoid ] shows that pressure and density in both plasmoids rise with time to values far in excess of the initial ones and that the rise is much stronger in the central plasmoid . for both plasmoids one can see that the density ratio @xmath109 remains much smaller than the corresponding pressure ratio
this is due to the strong plasma heating in the dissipation region around the x points . in the solar corona
one can expect that the size ratio of dissipation region and plasmoid is far smaller than in the simulation and that the pressure enhancement is to a larger fraction realized as a density enhancement .
the soft x - ray images of the plasmoid in the considered flare suggest a density ratio to the surrounding corona of @xmath110 ( ohyama & shibata @xcite ) . keeping in mind that the overall scales of the numerical experiment , @xmath85 and @xmath111 , are by orders of magnitude smaller than those of the flare , the formation of a plasmoid with a density enhancement @xmath112 in the flare by the secondary tearing and subsequent coalescence processes appears to be possible .
the central plasmoid can not move in the symmetrical simulation . in a small range of parameter values , it can tear and leave the box in two parts ( kliem & schumacher @xcite ) . if the condition of symmetry about the @xmath99 axis is relaxed ,
small asymmetries build up , which eventually lead to strong acceleration of the plasmoid .
previous simulations showed that a plasmoid can be formed also in the presence of asymmetries if their initial value is small ( schumacher & kliem @xcite ) .
there it was further found that the pressure and density in the plasmoid continue to rise during the acceleration and ejection of the plasmoid and that the velocity of the plasmoid can reach a significant fraction of the alfvn velocity ( @xmath113 ) .
as long as the plasmoid is being formed at a fixed position , its density is increasing .
only if the plasmoid starts to move along the current sheet to greater heights , it will expand and its pressure and density will decrease again according to the decreasing pressure of the surrounding medium .
the density _ ratio _ to the surrounding material may stay constant , or may even increase further , as long as the reconnection continues .
it is well known that particles are accelerated near the x points in the dc electric field associated with magnetic reconnection ( e.g. , deeg et al.@xcite ; moses et al .
@xcite ; birn et al .
a multiple x point reconnection process , characteristic of secondary tearing and other forms of impulsive bursty reconnection ( like the multiple coalescence instability ) , strongly enhances the number of accelerated particles and their maximum energy in comparison to single x point reconnection , where the particles can quickly escape ( ambrosiano et al.@xcite ; scholer & jamitzky @xcite ; kliem @xcite ; kliem et al .
peaks of the reconnection rate , such as seen in fig . [ r_rate ] , represent temporal peaks of the electric field at the x points and , hence , correspond to peaks of particle acceleration by the dc electric field .
the same is true for models of stochastic particle acceleration which suppose that the primary mhd wave spectrum is excited by the reconnection process ( miller et al
. @xcite ) .
although the physics of particle acceleration in reconnection regions still appears to be far from being understood , the existing investigations strongly suggest that a significant fraction of the particles in a dynamic current sheet can be accelerated to near - relativistic and relativistic energies .
an investigation of particle acceleration in a current sheet that undergoes secondary tearing is planned for a future study .
the strongest electric fields occur at the main magnetic x points adjacent to the plasmoid and , thus , a large fraction of the accelerated particles becomes trapped in the plasmoid . the other part can escape along the field lines of the reconnection outflow .
those particles escaping with the downward directed outflow are likely to become trapped in the magnetic cusp configuration between the current sheet and the underlying flare loop .
such trapping may explain the dominance of the trap - leakage component in the hard x - ray lightcurve and contribute to the decrease of the correlation between the general evolution of the hard x - ray and decimetric flux profiles in the course of the event ( i.e. , with increasing plasmoid height ) . in a strictly two - dimensional configuration ,
the particles in the central plasmoid are trapped indefinitely ( e.g. , scholer & jamitzky @xcite ) . in reality , also this part can escape from the plasmoid along the @xmath41 direction , either by ( slow ) guiding centre drifts into the @xmath41 direction , or by free streaming if a guide field @xmath114 exists .
it is not clear whether the dynamic current sheet that forms a plasmoid does also form a loop - shaped magnetic trap ( which is the canonical source model of several radio burst types , including decimetric continuum bursts ) .
a highly distorted magnetic loop , or flux rope , may be associated with the central plasmoid if a significant guide field component ( @xmath115 ) is present .
such a configuration was , in fact , suggested for the considered flare by ohyama & shibata ( @xcite ) .
however , due to the magnetic shear introduced by a @xmath114 in the vicinity of the plasmoid , the photospheric footpoints of the field lines that are connected to the plasmoid may be widely dispersed in this case , and it is not clear whether a loss - cone distribution function can be formed .
therefore , it is possible that an alternative source model is required for the radio burst . since the accelerated particles may gain excess perpendicular energy from drifts in the inhomogeneous magnetic field near the x points ( kliem @xcite ) , they may be immediately unstable ( i.e. , within the current sheet and the plasmoid ) to excitation of upper hybrid waves , which easily transform into radio emission , and the presence of a magnetic trap may not be necessary to form a radio source . in this section
we investigate how the scales found in the numerical experiment transform for parameters of the middle solar corona , where soft x - ray ejecta and decimetric radio sources are generally formed . as characteristic plasma parameters in the centre of the current sheet ,
we take the values that can be inferred for the flare of october 5 , 1992 from figs . 6 and 7 in ohyama & shibata ( @xcite ) , i.e. , @xmath27 @xmath11 and @xmath28 k ( sect . [ obs ] ) .
these values refer to a dense and heated ( i.e. , presumably active ) current sheet or possibly to the trailing part of the extended plasmoid , which had not yet reached a great height by the time of these parameter determinations ( 9:25:009:25:18 ut ) .
the plasma surrounding the current sheet was much less dense and much less hot , as was the current sheet itself in the yohkoh / sxt images at a later stage of the evolution .
a typical value of the coronal background plasma density at the considered heights ( @xmath116 cm ) is @xmath117 @xmath11 . as coronal background temperature
we will take @xmath118 k. the maximum instantaneous bandwidth of the decimetric radio burst suggests that the density varied by a factor @xmath119 within the radio source .
although the plasma parameters estimated by ohyama & shibata are subject to uncertainties that derive from necessary assumptions on the scale of the structures along the line of sight ( @xmath120 cm ) and the filling factor ( @xmath121 ) , they are consistent with the parameters derived for another flare which showed a plasma ejection ( tsuneta @xcite ) .
we note that a density of @xmath122 @xmath11 at the electron acceleration site was found also for other flares , using a different method ( by determining the starting frequency of forward and reversely drifting type iii burst pairs ; see aschwanden et al .
@xcite ) .
estimates of the magnetic field strength @xmath123 in the plasma surrounding the current sheet are also rather uncertain .
ohyama & shibata ( @xcite ) obtained @xmath124 g from the condition of pressure equilibrium across the current sheet , @xmath125 , using a rough estimate of the internal and external pressures , @xmath126 and @xmath127 , respectively ( actually they supposed a higher pressure than implied by the values of @xmath128 and @xmath129 given above ) .
this way of estimating @xmath123 is rather sensitive to the filling factor of the high - pressure current sheet plasma in the area of one detector pixel , @xmath130 for @xmath131 .
hence , a smaller as well as a higher value of the magnetic field would also be plausible .
nevertheless , we shall use @xmath124 g , which gave a reasonable estimate of the flare energy balance . using this with @xmath132 and @xmath133 , the alfvn velocity and
the plasma beta outside of the current sheet become @xmath134 cms@xmath5 and @xmath135 , respectively .
the current sheet half width can be estimated from the threshold condition for onset of kinetic current - driven plasma instabilities .
spontaneous excitation and re - excitation of anomalous resistivity implies that the ion - electron drift velocity in the sheet remains close to the critical value @xmath65 .
this critical drift velocity is of the order of the ion thermal velocity , @xmath136 , for the three most relevant instabilities and a wide range of parameter values .
the three most relevant instabilities are the lower - hybrid - drift instability , the electrostatic ion cyclotron instability , and the ion sound instability ( see , e.g. , the review by papadopoulos @xcite ) . using @xmath137 , ampere s law yields @xmath138 , where @xmath139 is the ion cyclotron radius . for our external values @xmath132 and @xmath133 we find @xmath140 cm .
the corresponding alfvn time is only @xmath141 s. when considering the direction along the magnetic field , we have to recall that a fluid model is valid only on scales @xmath142 larger than the mean free path , @xmath143 . in the fluid model ,
this length is the lower limit of inhomogeneity scales of @xmath53 , which set the distance @xmath144 between neighbouring x points . since the formation of inhomogeneities in the current sheet
is permanently enforced by the outflow along the @xmath90 axis of material that carries reconnected field lines , it appears reasonable to suppose that @xmath145 .
the mean free path is of order @xmath146 cm for both pairs of density and temperature values @xmath147 and @xmath148 [ but varies by a factor @xmath149 for the other combinations @xmath150 and @xmath151 . in the simulation ,
the time interval between main peaks of the reconnection rate is of order @xmath152 , which is about @xmath153 ( alfvn crossing times based on the average distance ( @xmath154 ) between neighbouring x points ) .
we now scale the timescale @xmath155 to the much larger distances expected in the corona , using @xmath145 ( which implies @xmath156 ) . for @xmath157 g and three different values of the density
, we derive the timescale @xmath158 in the range of temperatures @xmath159 k. this yields @xmath160 s for @xmath161 @xmath11 ; @xmath162 s for @xmath163 @xmath11 ; and @xmath164 s for @xmath165 @xmath11 . since the elongation and structuring of the area of anomalous resistivity ( which is the dominant factor controlling the dynamics ) takes place in the inflow sections of the current sheet , where the density and temperature are close to the external values @xmath132 and @xmath133 , the estimate @xmath166 s may be the most realistic one . in view of the uncertainties of the parameter determinations
, this is amazingly close to the observed intervals between the radio pulses , @xmath167 s. the poynting flux of electromagnetic energy toward the current sheet is @xmath168 , where @xmath169 is the vertical ( @xmath90- ) extent of the current sheet and the average inflow velocity can be expressed using the average reconnection rate @xmath170 as @xmath171 . using @xmath172 cm and @xmath173 ( from fig .
[ r_rate ] ) , we find @xmath174 ergs@xmath5 .
this appears to be a reasonable value , since it exceeds the increase rate of the thermal energy in the flare loop obtained by ohyama & shibata ( @xcite ) by a moderate factor ( @xmath175 ) consistent with the fact that not all of the inflowing electromagnetic energy can be dissipated .
the scaling @xmath176 shows that higher energy inflow rates can easily be achieved by modest increases of @xmath123 . on the other hand , the rate of particles flowing into the current sheet , @xmath177 , seems to fall short of the requirement on the acceleration rate of hxr - emitting particles , which is on the order of @xmath178 s@xmath5 for an m class flare .
our parameters @xmath132 , @xmath123 yield @xmath179 s@xmath5 , and the number of particles that can be accelerated out of this population must be significantly smaller .
the scaling @xmath180 shows that this dilemma can not be overcome without supposing very high values of @xmath132 and @xmath123 ( and @xmath133 to keep @xmath166 s ) and , hence , increasing @xmath181 by a huge amount .
even an increase of the average reconnection rate by one order of magnitude could only weaken but not eliminate the problem . whereas a sufficient number of particles is available within the current sheet to account for the decimetric radio source , it is likely that acceleration in additional volumes contributes significantly to the flux of hxr - emitting particles .
possibilities for this are ( 1 ) the spreading of the acceleration volume by waves which accelerate particles stochastically ( e.g. , miller et al .
@xcite ) , or ( 2 ) the interaction between the reconnection outflow and the low - lying loop , where the density is higher ( tsuneta @xcite ) , or ( 3 ) the transmission of magnetic stresses , which are created in the current sheet by reconnection , along the field lines to a more dense acceleration region near the base of the corona ( haerendel @xcite ) . the detailed correlation at small timescales between the hxr and the decimetric emissions
will then be relatively weak , at least for possibilities ( 2 ) and ( 3 ) , as is observed in the october 5 , 1992 flare and in many other events ( aschwanden et al .
@xcite ) .
maximum energies of particles accelerated by the dc electric field near x points have been found to reach @xmath182 ( in the ideal case of test particle behaviour with no back - reaction on the electromagnetic fields ) ( kliem @xcite ; kliem et al .
this falls into the mev range for all values of @xmath183 , @xmath184 , and @xmath185 considered here .
thus , energies of radio - emitting particles in flares ( @xmath186 kev ) can be reached easily .
the centre frequency of the decimetric continuum and pulses was drifting on average toward lower frequencies during the whole event .
this implies in our model that the plasmoid was not stationary but moving upwards during the whole radio burst so that the expansion at greater heights could counteract the density increase due to reconnection ( fig .
[ plasmoid ] ) . if we extrapolate the observed height - time plot of the plasmoid ( fig . 3 in ohyama & shibata @xcite ) backwards , based on the assumption of constant acceleration from zero velocity at @xmath187 to the observed velocity at 9:25 ut
, we arrive at a starting point , @xmath188 km , high above the top of the flare loop [ @xmath189 km ] , which is consistent with our simplified 2d model . since @xmath190 in the corona , the magnetic forces
dominate the acceleration of plasmoids . in our model of plasmoid formation in a preexisting current sheet , this implies that the reconnection rate at the x point below the plasmoid exceeded the reconnection rate at the x point above the plasmoid .
the imbalance of the reconnected flux then led to the upward bulk acceleration of the plasmoid .
it has been found in previous non - symmetric simulations of dynamic current sheet evolution ( using a smaller box and partly different initial conditions ; schumacher & kliem @xcite ) that the reconnection rate is most strongly enhanced during the buildup phase of a strong asymmetry , while the bulk velocity of the plasmoid reaches its peak value somewhat later ( by @xmath191 in those simulations ) .
it is thus expected that the most prominent temporary increases of the reconnection rate are associated with ( 1 ) enhanced particle acceleration , corresponding to enhanced radio and x - ray flux , ( 2 ) an initial rise of pressure and density in the plasmoid from the enhanced reconnection outflow into the plasmoid , corresponding to a rise of the frequency of plasma emission , and ( 3 ) the buildup of a strong asymmetry of the reconnected flux and , hence , a strong acceleration of the plasmoid , corresponding to a strong negative drift of the radio emission frequency in the inhomogeneous solar atmosphere , following the initial rise of the emission frequency .
the radio and hard x - ray data suggest such an association during the intervals of strong flux rise ( near 9:24:30 , 9:24:45 , and 9:25:50 ut ) .
the same effect is indicated also at the beginning of the event ( near 9:23:30 ut ) .
the interplay between the increase of the density in the plasmoid by reconnection and the decrease by upward motion and expansion can lead to a constant or even decreasing frequency drift rate of the radio emission in spite of increasing plasmoid velocity . also the soft x - ray observations suggest that the density of the plasmoid was only weakly decreasing during its rise from @xmath192 km at 9:25 ut to @xmath193 km at 9:26 ut .
ohyama & shibata ( @xcite ) derived densities @xmath194 @xmath11 and @xmath195 @xmath11 , respectively , at these two instants ( cf .
their tables 1 and 3 ) .
the very high frequency drift rates of the individual pulses were not measured .
an earlier statistical investigation of individual pulse drifts at decimetric frequencies in 10 pulsation events has shown a narrow distribution peaking at @xmath196 ghzs@xmath5 , consistent with compact sources of enhanced density and a scale height of @xmath197 km ( aschwanden & benz @xcite ) .
we note that the plasmoid observations during the october 5 , 1992 flare as well as our model are fully consistent with those source characteristics .
the weak statistical tendency of individual pulses toward positive frequency drift is consistent with the density increase of the plasmoid associated with each single coalescence event .
although the dynamical behaviour of the pulsating decimetric radio burst , as well as the majority of the hard and soft x - ray data , appear to be consistent with our model of the formation and acceleration of a ( single ) large - scale plasmoid by dynamic magnetic reconnection , more complex scenarios are not excluded .
for example , it is possible that the pulsating radio emission before 9:24:20 ut belonged to a separate plasmoid , which may have formed independently if the current sheet had a large extent along the line of sight ( @xmath41 direction in the simulation ) .
the drop of the radio flux around 9:24:10 ut may perhaps be explained by such a multiple source structure .
also the existence of two additional weak emission bands , which showed signs of pulsations not or only weakly correlated with the pulses of the main radio source ( fig . [ threebands ] ) , suggests the formation of two additional radio sources ( plasmoids separated from the main plasmoid in the @xmath90 or @xmath41 direction , which are not resolved in the soft x - ray images ) .
the variations of the frequency drift rate may have been influenced also by acceleration or deceleration of the plasmoid by the expanding loop , if it was in fact connected with the plasmoid .
the aspect ratio @xmath198 reaches @xmath199 in the simulation .
the estimates of sect . [ scales ] suggest that @xmath200 in the corona .
this suggests that the modulations of the reconnection rate , @xmath201 , become even deeper than in the simulation , since @xmath202 is reduced more strongly during the phases of sweet - parker reconnection .
this is in favour of deep modulations of the particle acceleration rate like that observed in the flare under study .
anomalous resistivity is created by kinetic current - driven instabilities , which have the ion cyclotron radius , @xmath203 , and the ion inertial length , @xmath204 , as natural scales . for our external plasma parameters @xmath123 , @xmath132 , and @xmath133 , for example
, we have @xmath205 cm and @xmath206 cm .
thus , it is probable that the kinetic processes lead to small - scale ( @xmath207 ) and short - lived ( @xmath208 ) inhomogeneities of the anomalous resistivity .
this may enforce intermittent excitation of small - scale magnetic islands .
such small - scale structures may be destroyed again by diffusion or may merge into larger islands and enter finally into the coalescence and plasmoid formation process on scales at which a fluid description is appropriate .
more complicated schemes of dynamic reconnection are conceivable , based on quasi - periodicities in complex multiple island coalescence processes ( kliem @xcite ) . in that paper
it was suggested that new islands are not formed as single entities , but that excitation of the tearing mode in the elongated flat sections of the current sheet creates a chain of islands .
the subsequent evolution may be rather similar to the dynamics obtained here . since the initial phase of the coalescence instability is an ideal mhd process , this instability develops faster than the tearing instability .
therefore , in the presence of perturbations , it takes over as the nonlinear evolution of the tearing instability ( inhibiting saturation of the tearing instability to a regular chain of islands ) .
the multiple coalescence of the island chain in a section of the current sheet leads to formation and amplification of plasmoids , similar to the evolution found here ; it may even be more dynamic .
cyclic repetition of the sequence of tearing instability , coalescence instability , and re - creation of long , flat current sheet sections between plasmoids may result in pulsed particle acceleration and radio emission .
some aspects of such a complex reconnection process have been studied by schumacher & kliem ( @xcite ) .
also externally triggered magnetic reconnection can show secondary tearing .
this was found by odstril & karlick ( @xcite ) for the case of triggering by a weak shock wave incident on a current sheet .
alternatively , or in addition to variable particle acceleration , the pulses of the radio flux may be caused by varying efficiency of the emission process in the disturbed and highly inhomogeneous plasma of the dynamic current sheet .
for example , the irregular oscillations of the central plasmoid , which were mentioned in sect . [ dynamic ] and are indicated by the variations of its peak pressure in fig .
[ plasmoid ] , could influence the radio emissivity .
however , varying particle acceleration is a natural and immediate consequence of dynamic reconnection .
a quantitative investigation of varying emission efficiency is made difficult by insufficient knowledge of the distribution function of the energetic particles and , hence , of the kinetic plasma physics of the emission process .
an interpretation of the weak anticorrelation between the decimetric pulses and the hxr lightcurve at short timescales ( fig .
[ cc ] ) in terms of our current sheet model is possible but remains speculative .
a tentative interpretation may be based on the oscillations of the central plasmoid : the outflow of plasma and accelerated particles from the main x points into the plasmoid is hampered , while the outflow away from the plasmoid is supported , during phases of plasmoid expansion in the @xmath90 direction ; and the situation is reversed during plasmoid expansion in the @xmath99 direction .
alternatively , the anticorrelation could , for example , be related to disturbances of the radio source caused by temporarily enhanced particle acceleration in other parts of the active region ( cf.sect .
[ scales ] ) .
studies of a larger sample of events will be required to confirm the anticorrelation as a general property and to determine possible causes .
a model in which the pulsating radio burst is associated with a plasmoid implies a strong role of trapping of downward propagating particles in the cusp between the current sheet and the underlying flare loop .
the weakness of pulsations in the hxr burst is thus naturally explained .
this would be difficult if the radio source were formed within the flare loop , as is commonly supposed for decimetric continuum bursts and was assumed for the october 5 , 1992 event by aschwanden & benz ( @xcite ) .
an association of radio continuum burst sources with plasmoid ejections had previously been found in case of isolated sources of moving type iv bursts at metric and decametric frequencies and much larger heights in the corona ( @xmath209 cm ; e.g. , gopalswamy et al .
recently , also hori ( @xcite ) suggested an association betwen drifting decimetric / metric continuum bursts and the ejection of plasmoids for the flares of october 5 , 1992 and june 28 , 1993 .
the proposed mechanism of variable particle acceleration may also underlie quasi - periodic sequences of type iii bursts ( aschwanden et al.@xcite ) .
these events are also associated with the impulsive phase of solar flares and their timescales are similar to those of the pulsations considered here ( @xmath2 s )
. a stronger role of particle escape and a weaker influence of particle trapping may lead to chains of type iii bursts instead of a pulsating radio continuum .
this appears possible in case of strongly asymmetric reconnection , where the field lines passing through the vicinity of the strongest x point below the plasmoid do not immediately close at the x point above the plasmoid , or in case of a sufficiently strong guide field component , @xmath115 , where the particles can easily escape from the plasmoid into the @xmath41 direction .
since type iii burst sources generally follow open field lines , the current sheet should extend into interplanetary space , i.e. , exist prior to the flare event in those cases .
\(1 ) we have presented a new model of pulsating radio bursts , in which the variations of the flux are caused by modulations of the particle acceleration in a highly dynamic reconnection process .
this process operates in extended coronal current sheets and includes ( in a fluid model ) the self - consistent evolution of anomalous resistivity .
it involves the repeated formation of magnetic islands , their coalescence , and the formation of one ( or several ) dense and hot plasmoid(s ) which can finally be ejected .
this model is able to explain the occurrence of irregular or quasi - periodic pulses with timescales in the range @xmath210 s. it is also consistent with the usually low amplitude of modulation of the hard x - ray emission and with the observed weak anticorrelation between the hard x - ray and decimetric emissions at short timescales in the flare of october 5 , 1992 considered here .
\(2 ) the unified explanation of plasmoid formation and pulsating radio emission supports the conclusion by ohyama & shibata ( @xcite ) that the current sheet in the considered flare was not formed by the ejection of the plasmoid ( contrary to the often favoured view of eruptive flare processes ) .
the model is consistent with both of the two remaining possibilities formation and acceleration of the plasmoid within a preexisting current sheet or simultaneous formation of both structures by a global mhd instability .
\(3 ) the emission mechanism of the radio waves may differ from the standard model of radio continuum bursts in the decimetric and metric range , which supposes particle injection and trapping in a stable coronal loop and subsequent evolution of a loss - cone particle distribution . instead
, a large fraction of the accelerated particles may only temporarily be trapped in the plasmoid , and the acceleration process itself may form an anisotropic velocity distribution , which is unstable against electrostatic or electromagnetic wave excitation .
\(4 ) we have confirmed and extended previous numerical studies of impulsive bursty reconnection , in particular the finding that magnetic reconnection , which is triggered by anomalous resistivity , proceeds in a highly variable manner if long current sheets are considered and the temporal and spatial evolution of the anomalous resistivity is self - consistently taken into account .
we gratefully acknowledge the very helpful comments by k .-
klein and the anonymous referee .
the zrich radio observations are partially financed by snf grant 20 - 53664.98
. technical support regarding the presentation of the ondejov radio data was given by h. mszrosov .
this work was supported by dlr grant 50oc9706 and by the key projects k1 - 003 - 601 and k1 - 043 - 601 , and the grant a3003707 of the academy of sciences of the czech republic .
the john von neumann - institut fr computing , jlich granted cray computer time .
ambrosiano j. , matthaeus w.h . ,
goldstein m.l .
, plante d. , 1988 , j. geophys .
93 , 14383 aschwanden m.j . , 1987 , solar phys .
111 , 113 aschwanden m.j .
, benz a.o . , 1986 ,
a&a 158 , 102 aschwanden m.j . ,
benz a.o . , 1995 ,
apj 438 , 997 aschwanden m.j .
, benz a.o . ,
kane s.r . , 1990 ,
a&a 229 , 206 aschwanden m.j . ,
benz a.o . ,
schwartz r.a . , 1993 ,
apj 417 , 790 aschwanden m.j . ,
benz a.o . , montello m.l . , 1994 ,
apj 431 , 432 aschwanden m.j . , bynum r.m . , kosugi t. , hudson h.s . , schwartz r.a
, 1997 , apj 487 , 936 aschwanden m.j . , kliem b. , schwarz u. , kurths j. , dennis b.r . , schwartz r.a . , 1998 ,
apj 505 , 941 bastian t.s . ,
benz a.o . , gary d.e . , 1998 , ara&a 36 , 131 birn j. , thomsen m.f . , borovsky j.e . , et al . , 1998 , j. geophys .
103 , 9235 biskamp d. , 1982 , z. naturforsch . 37a , 840 biskamp d. , 1986 , phys .
fluids 29 , 1520 biskamp d. , 1994 , phys . reports 237 , 179 deeg h .- j .
, borovsky j.e . , duric n. , 1991 , phys .
fluids b 3 , 2660 forbes t.g . , priest
e.r . , 1987 , rev .
25 , 1583 gopalswamy n. , kundu m.r .
, manoharan p.k . , et al . , 1997 ,
apj 486 , 1036 haerendel g. , 1994 , apjs 90 , 765 hori k. , 1999 , in : solar physics with radio observations , bastian t. , gopalswamy n. , shibasaki k. ( eds . ) , nobeyama radio observatory , nro report no
. 479 , p. 267 karlick m. , 1988 , bull . astron .
39 , 13 karpen j.t .
, antiochos s.k . ,
devore c.r . , 1995 , apj 450 , 422 kiplinger a.l . , dennis b.r .
, frost k.j . ,
orwig l.e .
, emslie a.g . , 1983 ,
apj 265 , l99 kliem b. , 1988 , esa sp-285 , vol . 2 , 117 kliem b. , 1994 , apjs 90 , 719 kliem b. , schumacher j. , 1996 , in : magnetodynamic phenomena in the solar atmosphere prototypes of stellar magnetic activity , uchida y. , kosugi t. , hudson h.s .
( eds . ) , kluwer , dordrecht , p. 305 kliem b. , schumacher j. , shklyar d.r .
, 1998 , adv .
space res .
21(4 ) , 563 krger a. , 1979 , introduction to solar radio astronomy and radio physics , d.reidel publ .
, dordrecht kuijpers j. , 1980 , in : radio physics of the sun , kundu m.r . , gergely t.e .
( eds . ) , iau symp .
86 , reidel , dordrecht , p. 341
kulsrud r.m .
, 1998 , phys .
plasmas 5 , 1599 magara t. , shibata k. , yokoyama t. , 1997 , apj 487 , 437 masuda s. , kosugi t. , hara h. , tsuneta s. , ogawara y. , 1994 , nature 371 , 495 miller j.a .
, larosa t.n . ,
moore r.l . , 1996 ,
apj 461 , 445 moses r.w . , finn j.m . , ling k.m . , 1993 , j. geophys .
98 , 4013 odstril d. , karlick m. , 1997 , a&a 326 , 1252 ohyama m. , shibata k. , 1998 , apj 499 , 934 papadopoulos k. , 1979 , in : dynamics of the magnetosphere , akasofu s .- i .
( ed . ) , reidel , dordrecht , p. 289
priest e.r .
, 1982 , solar magnetohydrodynamics , d.reidel publ .
, dordrecht priest e.r .
, 1985 , rep .
48 , 955 priest e.r .
, forbes t.g .
, 1986 , j. geophys .
91 , 5579 sato j. , sawa m. , masuda s. , et al . , 1998 , the yohkoh hxt image catalogue , october 1991august 1998 , nobeyama radio observatory , nobeyama sato t. , hayashi t. , 1979 , phys
. fluids 22 , 1189 scholer m. , 1991 , geophys .
fluid dyn .
62 , 51 scholer m. , jamitzky f. , 1987 , j. geophys .
92 , 12181 scholer m. , roth d. , 1987 , j. geophys .
92 , 3223 schumacher j. , kliem b. , 1996 , phys .
plasmas 3 , 4703 schumacher j. , kliem b. , 1997 , phys .
plasmas 4 , 3533 schumacher j. , kliem b. , seehafer n. , 2000 , phys .
plasmas 7 , 108 tsuneta s. , 1995 , publ
. astron . soc .
691 tsuneta s. , 1996 , apj 456 , 840 ( erratum : apj 464 , 1055 ) tsuneta s. , 1997 , apj 483 , 507 ugai m. , 1992 , phys .
fluids b4 , 2953 ugai m. , 1993 , phys .
fluids b5 , 3021 ugai m. , 1995 , phys .
plasmas 2 , 388 ugai m. , tsuda t. , 1977 , j. plasma phys .
17 , 337 yokoyama t. , shibata k. , 1994 , apj 436 , l197 | decimetric radio observations of the impulsive solar flare on october 5 , 1992 , 09:25 ut show a long series of quasi - periodic pulsations deeply modulating a continuum in the 0.62 ghz range that is slowly drifting toward lower frequencies .
we propose a model in which the pulsations of the radio flux are caused by quasi - periodic particle acceleration episodes that result from a dynamic phase of magnetic reconnection in a large - scale current sheet .
the reconnection is dominated by repeated formation and subsequent coalescence of magnetic islands ( known as `` secondary tearing '' or `` impulsive bursty '' regime of reconnection ) , while a continuously growing plasmoid is fed by newly coalescing islands .
such a model , involving a current sheet and a growing plasmoid , is consistent with the yohkoh observations of the same flare ( ohyama & shibata @xcite ) .
we present two - dimensional mhd simulations of dynamic magnetic reconnection that support the model . within the framework of the proposed interpretation ,
the radio observations reveal details of plasmoid formation in flares . |
the inclusive production of a single charged hadron , @xmath3 , in the annihilation process @xmath4 has been measured at many different @xmath0 colliders over a wide range of center - of - mass ( cm ) energies , @xmath5 , between 3 and 183 gev @xcite . here
, @xmath3 may either refer to a specific hadron species , _
e.g. _ , pion , kaon , or proton , or to the sum over all charged - hadron species .
a large amount of precise data has now become available from various experimental collaborations at cern lep1 and slac slc , who started taking data several years ago . the process ( [ process ] ) is particularly suitable in order to study the fragmentation of quarks and gluons into hadrons .
the information contained in fixed - target , hadron - collider , and @xmath6-scattering data is less useful , since it is obscured by theoretical uncertainties from the parton distribution functions and the choice of factorization scales connected with the initial states .
the partonic cross sections pertinent to process ( [ process ] ) can be entirely calculated in perturbative qcd with no additional input , except for the strong - coupling constant @xmath7 .
they are known at next - to - leading order ( nlo ) @xcite and even at next - to - next - to - leading order ( nnlo ) @xcite .
the subsequent transition of the partons into hadrons takes place at an energy scale of the order of 1 gev and can , therefore , not be treated in perturbation theory . instead
, the hadronization of the partons is described by fragmentation functions ( ffs ) , @xmath8 .
their values correspond to the probability that the parton @xmath9 , which is produced at short distance , of order @xmath10 , fragments into the hadron @xmath3 carrying the fraction @xmath11 of the momentum of @xmath9 . in the case of process ( [ process ] )
, @xmath12 is typically of order @xmath13 .
given their @xmath11 dependence at some scale @xmath14 , the evolution of the ffs with @xmath12 may be computed perturbatively . at present ,
the evolution equations are only known through nlo .
consistency requires that we do not include the nnlo corrections to the partonic cross sections @xcite in our analysis of process ( [ process ] ) via unpolarized photons and @xmath1 bosons .
the theoretical analysis of the fragmentation process in @xmath0 annihilation is of interest for several reasons .
first , it allows us to test qcd quantitatively within one experiment observing hadrons at different values of @xmath5 and to determine the strong - coupling constant , @xmath15 , to be compared with determinations from other observables and/or processes .
second , since according to the factorization theorem the ffs are independent of the process in which they have been determined , they can be used for quantitative predictions of inclusive single hadron cross sections in other processes , like @xmath16 , @xmath6 , @xmath17 , and @xmath18 scattering .
third , the exact knowledge of the ffs , in particular for identified hadrons , may help to elucidate the fundamental features of hadronization and to constrain models used for the calculation of complete final states in various high - energy reactions .
after the pioneering leading - order ( lo ) analyses of pion , kaon @xcite , and charmed - meson @xcite ffs in the late 70s , there had long been no progress on the theoretical side of this field . in the mid 90 s , nlo ff sets for @xmath19 @xcite , @xmath20 , @xmath21 , and @xmath22 mesons @xcite have been constructed through fits to data of @xmath0 annihilation , mostly generated with monte carlo event generators . in 1994/95 , two of us , together with binnewies , determined the ffs of quarks and gluons into @xmath23 and @xmath24 mesons at lo and nlo from genuine experimental data , adopting two different strategies @xcite . in the first analysis @xcite , we included in the fit only the data on @xmath23 and @xmath24
production taken by the tpc / two - gamma collaboration @xcite at slac pep , with energy @xmath25 gev .
these data combine small statistical errors with fine binning in @xmath11 and are thus more constraining than data collected by other experiments in the energy range from 5.2 gev to 44 gev .
charged - pion and -kaon data at other energies were only used for cross checks . furthermore , theoretical predictions obtained using these ffs were compared with data on the inclusive production of unidentified charged hadrons taken at the pep , desy petra , and kek tristan colliders making an assumption on the @xmath26 contribution based on information extracted from the tpc data @xcite . in 1994 ,
when the first analysis @xcite was performed , data with identified quark flavours did not exist .
therefore , we had to make the assumption that the @xmath13 , @xmath27 , and @xmath28 ( @xmath29 , @xmath27 , and @xmath28 ) quarks fragment into @xmath20 ( @xmath21 ) mesons in the same way . in our second analysis @xcite , we could remove this assumption because new data on charged - hadron production with partial flavour separation had been released by the aleph collaboration @xcite at lep1 ( see also refs .
apart from the usual data sample without flavour separation , they also presented light- and @xmath28-quark - enriched samples . here and in the following , we consider @xmath30 , @xmath29 , and @xmath13 quarks as light .
in addition , we included in our fits accurate @xmath20 and @xmath21 data from aleph @xcite and information on tagged three - jet events from opal @xcite at lep1 , which constrained the gluon ffs . in order to describe the production of charged hadrons in terms of the @xmath20 and @xmath21 ffs
, we needed some information on the fragmentation of quarks and gluons into protons and antiprotons . to that end
, we introduced a function @xmath31 which parameterizes the ratio of the @xmath32 and @xmath20 production cross sections as measured by aleph @xcite .
it is clear that this analysis @xcite suffered from the lack of specific data on the fragmentation of tagged quarks and gluons to @xmath20 , @xmath21 , and @xmath26 hadrons , rather than just to charged hadrons of unidentified species .
this lack has been remedied in 1998 by the advent of new data from delphi @xcite at lep1 and sld @xcite at slc .
besides other things , sld @xcite measured the differential production cross section , as a function of the scaled momentum @xmath33 , for the identified hadron species @xmath20 , @xmath21 , and @xmath26 in @xmath1-boson decays to light , @xmath27 , and @xmath28 quarks and in such decays without flavour separation . similarly , delphi @xcite presented @xmath20 , @xmath21 , and @xmath26 production data , covering the full momentum range up to 45.6 gev , discriminating between @xmath34 , @xmath35 , and @xmath36 events
. there is also new specific information that allows us to better control the gluon ff .
in fact , new gluon - tagged three - jet data of inclusive charged - hadron production was released by aleph @xcite and opal @xcite . in the opal @xcite analysis
, the gluon jets were identified using the method proposed by gary @xcite .
this method is based on rare events of the type @xmath37 in which the @xmath38- and @xmath39-quark jets appear in the same hemisphere of a multihadronic @xmath0 annihilation event .
the object @xmath40 , which is taken to be the gluon jet , is defined by all hadrons observed in the hemisphere opposite to that containing the @xmath38- and @xmath39-quark jets .
this method allows for the extraction of the gluon ff with as little theoretical bias as possible .
it is supposed to be superior to the earlier opal @xcite measurement of the gluon ff based on a jet - finding algorithm to define the gluon jets , which entered the bkk analysis @xcite .
the new opal @xcite data set comprises more points than the previous one @xcite , and the energy of the gluon jets is now well defined , with a mean value of @xmath41 gev .
it is the purpose of this work to make full use of all the recent data from aleph @xcite , delphi @xcite , opal @xcite , and sld @xcite in order to construct new sets of lo and nlo ffs without imposing relations between the ffs of different flavours , except that we identify the ffs of @xmath30 and @xmath29 quarks into @xmath20 mesons and those of @xmath30 and @xmath13 quarks into @xmath21 mesons , as in refs .
@xcite . in the case of protons and antiprotons
, we fix the @xmath29-quark ffs to be half of the @xmath30-quark ffs . for completeness , we also include in our fits the older @xmath20 , @xmath21 , and @xmath32 production data without flavour separation from aleph @xcite , which also entered the bkk analysis @xcite . in order to be sensitive to the running of @xmath7 , _
i.e. _ , to be able to fit the asymptotic scale parameter , @xmath42 , we also include the tpc data @xcite , as in refs .
there is also precise data on charged - hadron production with flavour separation from aleph @xcite , who provided light- , @xmath27- , and @xmath28-quark - enriched samples , and from opal @xcite , who provided light- and @xmath28-quark - enriched samples . to be on the conservative side
, we exclude this data from our fits because the contribution due to charged hadrons other than @xmath20 , @xmath21 , and @xmath32 is unknown and may be comparable or larger than the experimental errors .
in fact , it turns out that this data yields rather sizeable @xmath43 values relative to the theoretical predictions obtained from our ffs
. we also test our ffs against data from @xmath0 colliders with lower cm energies , namely , from desy doris @xcite and petra @xcite .
furthermore , we study the longitudinal cross section for the production of charged hadrons on the @xmath1-boson resonance and compare the results with data from aleph @xcite , delphi @xcite , and opal @xcite .
this work is organized as follows . in section 2 , we summarize the data
included in our fits , outline the theoretical framework , describe our fitting procedure , present our ffs , and discuss their goodness .
furthermore , we check our ffs against @xmath0 data which are not included in our fits . in section 3 , we investigate for charged - hadron production how well our ffs satisfy the momentum sum rule , which we did not impose on our fits .
furthermore , we examine to what extent our ffs also account for the longitudinal cross section of charged - hadron production on the @xmath1-boson resonance , and how our nlo gluon ff compares with the one of the bkk set @xcite and with the one recently measured by delphi @xcite .
our conclusions are summarized in section 4 .
.cm energies , types of data , and @xmath44 values obtained at lo and nlo for various data samples . samples not used in the fits are marked by asterisks . [
cols="^,<,<,<,<,<,<,<",options="header " , ] the comparisons presented in fig .
10 provide us with a very useful check of the gluon ff in the low-@xmath11 range , where the data have small errors .
in particular , the good agreement in the case of flavour separation nicely confirms the relative importance of the light- and @xmath28-quark ffs relative to the gluon ff .
since the @xmath44 values obtained at nlo for the longitudinal cross section are comparable to those achieved in our nlo fit , we do not expect that the inclusion of the information on the longitudinal cross section in our nlo fit would significantly change the outcome .
as is well known , the gluon ff only enters the prediction for @xmath45 at @xmath46 , while at @xmath47 it only contributes indirectly via the @xmath12 evolution .
therefore , we have used the aleph @xcite and opal @xcite data on gluon - tagged three - jet events in order to constrain it .
these data were already compared with the fitted gluon ff in fig .
it is of interest to compare this gluon ff with the bkk @xcite one and the one recently extracted by delphi @xcite from their three - jet events for scales in the range @xmath48 gev . in ref .
@xcite , a parameterization , valid for this @xmath49 range and for @xmath50 , was generated adopting an ansatz similar to the one introduced in ref .
@xcite . in fig .
11 , the @xmath11 dependences of these three gluon ffs are studied for @xmath51 , 52.4 , 80.2 , and 200 gev .
the results for @xmath52 and 80.2 gev are compared with the aleph @xcite and opal @xcite gluon - tagged three - jet data , which are included in our fits . from fig .
11 the following observations can be made .
our new gluon ff is rather similar to the bkk one , especially for larger values of @xmath49 . at @xmath53 gev ,
the delphi gluon ff is somewhat steeper than the new and bkk ones , except for large @xmath11 values , and it agrees slightly better with the opal @xcite data . on the other hand , our new gluon ff agrees best with the aleph @xcite data at @xmath52 gev . at @xmath51 gev
, the delphi gluon ff has a flatter @xmath11 dependence than the other ones and overshoots them , while at @xmath54 gev the situation is similar to the case of @xmath53 gev .
the lep1 and slc experiments have provided us with a wealth of high - precision experimental information on how partons fragment into low - mass charged hadrons @xcite .
the data partly come as light- , @xmath27- , and @xmath28-quark - enriched samples without @xcite or with identified final - state hadrons ( @xmath20 , @xmath21 , and @xmath26 ) @xcite or as gluon - tagged three - jet samples without hadron identification @xcite .
this new situation motivated us to update , refine , and extend the bkk analysis @xcite by generating new lo and nlo sets of @xmath20 , @xmath21 , and @xmath26 ffs .
the @xmath11 distributions of the resulting ffs at their starting scales @xmath55 are given by eq .
( [ ansatz ] ) with the parameters listed in table 2 .
the @xmath12 evolution is determined by the timelike altarelli - parisi equations in the respective orders , lo or nlo , which are summarized in the appendix of ref .
@xcite . should be replaced by @xmath56 . ] the evolution procedure in mellin space is described in ref .
@xcite . a fortran subroutine which returns the values of the @xmath8 functions for given values of @xmath11 and @xmath12
may be downloaded from the url http://www.desy.de/~poetter/kkp.html or obtained upon request from the authors . by also including in our fits @xmath20 , @xmath21 , and @xmath26 data ( without flavour separation ) from pep
@xcite , with cm energy @xmath2 gev , we obtained a handle on the scaling violation in the fragmentation process , which allowed us to extract lo and nlo values of @xmath57 .
we found @xmath58 at lo and @xmath59 at nlo , where the errors are experimental .
these results are in perfect agreement with what the particle data group currently quotes as the world average , @xmath60 @xcite .
our strategy was to only include in our fits lep1 and slc data with both flavour separation and hadron identification @xcite , gluon - tagged three - jet samples with a fixed gluon - jet energy @xcite , and the @xmath20 , @xmath21 , and @xmath26 data sets from the pre - lep1/slc era with the highest statistics and the finest binning in @xmath11 @xcite .
other data served us for cross checks .
in particular , we probed the scaling violation through comparisons with @xmath20 , @xmath21 , and @xmath26 data from doris and petra , with cm energies between 5.4 and 34 gev @xcite . furthermore , we tested the gluon ff , which enters the unpolarized cross section only at nlo , by comparing our predictions for the longitudinal cross section , where it already enters at lo , with available data @xcite . finally , we directly compared our gluon ff with the one recently measured by delphi in three - jet production with gluon identification as a function of @xmath11 at various scales @xmath12 @xcite .
all these comparisons led to rather encouraging results .
we also verified that our ffs satisfy reasonably well the momentum sum rules , which we did not impose as constraints on our fits .
all these cross checks make us believe that our ffs should allow for a reliable description of inclusive charged - hadron production in all kinds of high - energy - collision experiments over wide ranges of @xmath11 and @xmath12
. the very - high-@xmath12 regime will be accessible with the cern lhc and a future @xmath0 linear collider .
our ffs are bound to break down at very low values of @xmath11 , where it becomes necessary to modify the ansatz ( [ ansatz ] ) according to the mlla @xcite .
this is beyond the scope of the present analysis , but should eventually be investigated @xcite . after the submission of this paper , a preprint @xcite appeared which also presents nlo ffs for @xmath20 , @xmath21 , and charged hadrons .
these ffs are fitted to the aleph @xcite , sld @xcite , and tpc @xcite data assuming certain power laws in the light - quark sector .
institut fr theoretische physik is supported in part by the deutsche forschungsgemeinschaft through grant no .
kn 365/1 - 1 , by the bundesministerium fr bildung und forschung through grant no .
05 ht9gua 3 , and by the european commission through the research training network _
quantum chromodynamics and the deep structure of elementary particles _ under contract no .
erbfmrx - ct98 - 0194 .
v. barger , t. gottschalk , r.j.n .
phillips , phys .
70 b ( 1977 ) 51 ; + j.f .
owens , e. reya , m. glck , phys .
d 18 ( 1978 ) 1501 ; + r. baier , j. engels , b. petersson , z. phys .
c 2 ( 1979 ) 265 ; + m. anselmino , p. kroll , e. leader , z. phys .
c 18 ( 1983 ) 307 .
cowan , _ in _ proceedings of the xxvii international conference on high energy physics , glasgow , scotland , uk , 2027 july 1994 , edited by p.j .
bussey and i.g .
knowles ( institute of physics publishing , bristol , 1995 ) p. 883 .
argus collaboration , h. albrecht et al . , z. phys .
c 44 ( 1989 ) 547 ; + tasso collaboration , w. braunschweig et al . , z. phys .
c 47 ( 1990 ) 167 ; + aleph collaboration , d. buskulic et al .
, z. phys .
c 64 ( 1994 ) 361 ; + opal collaboration , r. akers et al .
, z. phys .
c 67 ( 1995 ) 389 . | we present new sets of fragmentation functions for charged pions , charged kaons , and protons , both at the leading and next - to - leading orders .
they are fitted to the scaled - momentum distributions of these hadrons measured in @xmath0 annihilation on the @xmath1-boson resonance at cern lep1 and slac slc .
these data partly come as light- , charm- , bottom - quark - enriched and gluon - jet samples , which allows us to treat all partons independently , after imposing the su(2 ) flavour symmetry relations . in order to gain sensitivity to the scaling violation in fragmentation , we also include data from slac pep , with center - of - mass energy @xmath2 gev , in our fits .
this allows us to also determine the strong - coupling constant , with a competitive error .
lep1 data on the longitudinal cross section as well as desy doris and petra data at lower energies nicely agree with theoretical predictions based on our fragmentation functions .
pacs numbers : 13.65.+i , 13.85.ni , 13.87.fh 16.cm 22.cm @=11 tempcntc citex[#1]#2@fileswauxout tempcnta@tempcntb@neciteaciteforciteb:=#2citeo#1 citeotempcnta > tempcntbciteacitea , tempcnta = tempcntbtempcnta |
random matrices arise in many mathematical contexts , and it is natural to ask about the properties that such matrices satisfy . if we choose a matrix with integer entries at random , for example , we would like to know the probability that it has a particular integer as an eigenvalue , or an integer eigenvalue at all . similarly ,
if we choose a matrix with real entries at random , we would like to know the probability that it has a real eigenvalue in a particular interval .
certainly the answer depends on the probability distribution from which the matrix entries are drawn . in this paper , we are primarily concerned with uniform distribution , so for both integer - valued and real - valued cases we must restrict the entries to a bounded interval . in an earlier paper @xcite , the authors show that random @xmath1 matrices of integers almost never have integer eigenvalues .
an explicit calculation by hetzel , liew , and morrison @xcite shows that a @xmath0 matrix with entries independently chosen uniformly from @xmath2 $ ] has real eigenvalues with probability @xmath3 .
this calculation gives hope that our more precise questions about eigenvalues of a particular size might be accessible in the @xmath0 setting .
the purpose of this paper is to resolve these questions , once we make them suitably precise . for an integer @xmath4 ,
let @xmath5 denote the uniform probability space of @xmath0 matrices of integers with absolute value at most @xmath6 .
note that @xmath7 .
we will obtain exact asymptotics for the number of matrices in @xmath5 having integer eigenvalues and , more precisely , for the number of matrices with a given integer eigenvalue @xmath8 . for any integer @xmath8 , define @xmath9 and let @xmath10 define the function @xmath11\to{{\mathbb r}}$ ] by @xmath12 and @xmath13 ( where @xmath14 is the natural logarithm ) . then for any integer @xmath8 between @xmath15 and @xmath16 , @xmath17 where the implied constant is absolute . on the other hand , if @xmath18 then @xmath19 is empty .
[ theorem : particular integer eigenvalue ] we remark that the function @xmath20 is continuous and , with the exception of the points of infinite slope at @xmath21 , differentiable everywhere ( even at @xmath22 , if we imagine that @xmath20 is defined to be 0 when @xmath23 ) .
notice that equation is technically not an asymptotic formula when @xmath8 is extremely close to @xmath24 , because then the value of @xmath25 can have order of magnitude @xmath26 or smaller , making the `` main term '' no bigger than the error term . however , equation is truly an asymptotic formula for @xmath27 , where @xmath28 is any function tending to infinity ( the exponent @xmath29 arises because @xmath20 approaches 0 cubically as @xmath30 tends to 2 from below ) . by summing the formula over all possible values of @xmath8
, we obtain an asymptotic formula for @xmath31 .
we defer the details of the proof to section [ enumeration ] .
let @xmath32 .
the probability that a randomly chosen matrix in @xmath5 has integer eigenvalues is asymptotically @xmath33 .
more precisely , @xmath34 [ corollary : all integer eigenvalues ] if @xmath35 has eigenvalue @xmath8 , then the scaled matrix @xmath36 has eigenvalue @xmath37 , which is the argument of @xmath38 that appears on the right - hand side of .
thus one interpretation of theorem [ theorem : particular integer eigenvalue ] is that for large @xmath6 , the rational eigenvalues of @xmath36 tend to be distributed like the function @xmath38 .
note that the entries of @xmath36 are sampled uniformly from a discrete , evenly - spaced subset of @xmath2 $ ] . as @xmath39
this probability distribution converges in law to the uniform distribution on the interval @xmath2 $ ] .
let @xmath40)}$ ] denote the probability space of all @xmath0 matrices whose entries are independent random variables drawn from this distribution .
one might ask whether the distribution given by theorem [ theorem : particular integer eigenvalue ] is just a discrete approximation to the distribution of eigenvalues in @xmath40)}$ ] ; the answer , perhaps surprisingly , is no .
the next theorem provides this latter distribution .
define @xmath41 to be the density function for real eigenvalues of matrices in @xmath40)}$ ] : if @xmath42 is a randomly chosen matrix from @xmath40)}$ ] , then the expected number of eigenvalues of @xmath42 in the interval @xmath43 $ ] is @xmath44 . then @xmath45 and @xmath46 [ theorem : distribution of real eigenvalues ] as in the case of @xmath20 ,
the function @xmath41 is continuous and differentiable everywhere , with the exception of the points of infinite slope at @xmath21 .
( the value @xmath47 makes the function continuous there , although the value of a density function at a single point is irrelevant to the probability distribution . )
it also shares the same cubic decay as @xmath30 tends to 2 from below . however , there are obvious qualitative differences between the functions @xmath38 and @xmath48 . in figure [ figure ] we plot @xmath38 and @xmath48 on the same axes , normalized so that the area under each is 2 ( these normalized versions are denoted @xmath49 and @xmath50 in our earlier paper @xcite ) . in the case of @xmath5 , this normalization corresponds to conditioning on having integer eigenvalues , that is , scaling by the probability @xmath33 from corollary [ corollary : all integer eigenvalues ] . for @xmath40)}$ ]
we are conditioning on having real eigenvalues , which occurs with probability @xmath3 ( this can be obtained by integrating @xmath41 , analogously to the proof of corollary [ corollary : all integer eigenvalues ] ; the computation by hetzel , liew , and morrison @xcite is more direct ) .
( @xmath38 normalized ) versus @xmath50 ( @xmath48 normalized),width=384 ] note that the distribution @xmath41 is bimodal , having its maxima at @xmath51 thus , a random matrix in @xmath40)}$ ] is more likely to have an eigenvalue of magnitude near @xmath52 than one of magnitude near @xmath53 .
we expect this would still hold if we were to condition on matrices in @xmath40)}$ ] having rational eigenvalues , since any matrix with real eigenvalues is a small perturbation from one with rational eigenvalues . that this is not true for @xmath20 shows that the eigenvalue distribution of @xmath5 is not purely the result of magnitude considerations but
also encodes some of the arithmetic structure of the integers up to @xmath6 .
we remark that theorem [ theorem : particular integer eigenvalue ] can also be obtained from a powerful result of katznelson @xcite .
let @xmath54 be a convex body containing the origin in @xmath55 , and embed the set of @xmath56 integer matrices @xmath57 as lattice points @xmath58 .
then theorem 1 of @xcite gives an asymptotic formula for the number of singular integer matrices inside the dilate @xmath59 .
taking @xmath60 ^ 4 $ ] then yields an asymptotic formula for @xmath61 , and more generally one can obtain @xmath62 by adding and subtracting appropriate shifts of @xmath54 .
the asymptotic formula in @xcite is defined in terms of an unusual singular measure supported on the zariski - closed subset of @xmath55 corresponding to singular matrices .
the explicit computation of this measure is roughly analogous to our case - by - case considerations in section [ section : key proof ] , modulo the significant complications of carrying error terms .
our techniques are more elementary , but katznelson s results apply in theory to matrices of any size , whereas our methods become unwieldy even for @xmath63 matrices . in the case of @xmath1 matrices with entries independently chosen from a gaussian distribution
, a great deal more is known .
edelman @xcite has computed the exact distributions of the real and complex eigenvalues for any @xmath64 , as well as the number of real eigenvalues ( for instance , the probability of having all @xmath64 eigenvalues real is precisely @xmath65 ) . as @xmath66
, the real eigenvalues , suitably rescaled by a factor of @xmath67 , converge to the uniform distribution on @xmath2 $ ] .
similarly , the complex eigenvalues converge to the `` circular law '' predicted by girko @xcite , namely the uniform distribution on the unit disk centered at the origin .
very recently , tao and vu @xcite have shown that the circular law is universal : one can replace the gaussian distribution by an arbitrary distribution with mean 0 and variance 1 .
similar results have been established for random symmetric matrices with entries independently chosen from a gaussian distribution ( the `` wigner law '' ) or from other distributions . those who are interested in the connections between analytic number theory and random matrix theory might wonder whether those connections are related to the present paper .
the matrices in that context , however , are selected from classical matrix groups , such as the group of @xmath1 hermitian matrices , randomly according to the haar measures on the groups .
the relationship to our results is therefore minimal .
we begin with some elementary observations about @xmath56 matrices that will simplify our computations .
the first lemma explains why the functions @xmath38 and @xmath48 are supported only on @xmath68 $ ] .
any eigenvalue of a matrix in @xmath5 is bounded in absolute value by @xmath16 .
any eigenvalue of a matrix in @xmath40)}$ ] is bounded in absolute value by 2 .
[ lemma : baby gershgorin ] we invoke gershgorin s `` circle theorem ''
@xcite , a standard result in spectral theory : let @xmath69 be an @xmath1 matrix , and let @xmath70 denote the disk of radius @xmath71 around the complex number @xmath72 .
then gershgorin s theorem says that all of the eigenvalues of @xmath42 must lie in the union of the disks @xmath73 in particular , if all of the entries of @xmath42 are bounded in absolute value by @xmath74 , then all the eigenvalues are bounded in absolute value by @xmath75 .
the key to the precise enumeration of @xmath76 is the simple structure of singular integer matrices : for any singular matrix @xmath77 , either at least two entries of @xmath42 equal zero , or else there exist nonzero integers @xmath78 with @xmath79 such that @xmath80 moreover , this representation of @xmath42 is unique up to replacing each of @xmath78 by its negative . [ baby.enumeration ] if one of the entries of @xmath42 equals zero , then a second one must equal zero as well for the determinant to vanish .
otherwise , given @xmath81 with none of the @xmath82 equal to zero , define @xmath83 , and set @xmath84 and @xmath85 , so that @xmath79 . since @xmath42 is singular , the second row of @xmath86 must be a multiple of the first row that is , there exists a real number @xmath87 such that @xmath88 and @xmath89 .
since @xmath90 and @xmath91 are relatively prime , moreover , @xmath87 must in fact be an integer .
this argument shows that every such matrix has one such representation .
if @xmath92 is another such representation , then @xmath93 implies @xmath94 , which shows that @xmath95 ; the equalities @xmath96 , @xmath97 , and @xmath98 follow quickly . for a @xmath56 matrix @xmath99
we define @xmath100 it is easily seen that this is the discriminant of the characteristic polynomial of @xmath42 .
we record the following elementary facts , which will be useful in the proof of lemma [ lemma : double eigenvalue ] and proposition [ f_w proposition ] .
let @xmath42 be a @xmath101 matrix with real entries .
a. @xmath42 has repeated eigenvalues if and only if @xmath102 . b. @xmath42 has real eigenvalues if and only if @xmath103 . c. @xmath104 if and only if @xmath42 has two real eigenvalues of opposite sign .
d. if @xmath105 and @xmath103 , then the eigenvalues of @xmath42 have the same sign as @xmath106 . [ discriminant ]
let @xmath107 denote the eigenvalues of @xmath42 , so that @xmath108 , @xmath109 , and @xmath110 , each of which is real . parts ( a ) , ( b ) and ( d ) follow immediately from these observations , and part ( c ) from the fact that @xmath111 if @xmath107 are complex .
the next lemma gives a bound for the probability of a matrix having repeated eigenvalues .
it is natural to expect this probability to converge to 0 as @xmath6 increases , and indeed such a result was obtained in @xcite for matrices of arbitrary size .
we give a simple proof of a stronger bound for the @xmath56 case , as well as an analogous qualitative statement for real matrices which will be helpful in the proof of theorem [ theorem : distribution of real eigenvalues ] . the number of matrices in @xmath5 with a repeated eigenvalue is @xmath112 for every @xmath113 .
the probability that a random matrix in @xmath40)}$ ] has a repeated eigenvalue or is singular is @xmath53 .
[ lemma : double eigenvalue ] by lemma [ discriminant](a ) , the @xmath0 matrix @xmath114 has a double eigenvalue if and only if @xmath115 . for matrices in @xmath40)}$ ]
this is easily seen to be a zero - probability event , as is the event that @xmath116 . for matrices in @xmath5 , we enumerate how many can satisfy @xmath115 . if @xmath117 then there are @xmath118 choices for @xmath119 ; otherwise there are at most @xmath120 choices if @xmath121 and no choices otherwise .
( here @xmath122 is the number - of - divisors function ; the factor of 2 comes from the fact that @xmath91 and @xmath123 can be positive or negative , while the `` at most '' is due to the fact that not all factorizations of @xmath124 result in two factors not exceeding @xmath6 . )
therefore the number of matrices in @xmath5 with a repeated eigenvalue is at most @xmath125 where the inequality follows from @xmath126 and the well - known fact that @xmath127 for any @xmath113 ( see for instance @xcite ) .
let @xmath128 be the mbius function , characterized by the identity @xmath129 the well - known dirichlet series identity @xmath130 is valid for @xmath131 ( see , for example , ( * ? ? ?
* corollary 1.10 ) ) .
in particular , @xmath132 , and we can estimate the tail of this series ( using @xmath133 ) to obtain the quantitative estimate @xmath134 for nonzero integers @xmath90 , @xmath91 and parameters @xmath6 , @xmath8 , define the function @xmath135 then @xmath136 where the implied constant is independent of @xmath8 and @xmath6 .
[ initial count and mu insertion ] fix an integer @xmath137 , and let @xmath138 , so that @xmath139 is singular . by lemma [ baby.enumeration ] , either at least two entries of @xmath139 equal zero , or else @xmath139 has exactly two representations of the form . in the former case ,
there are @xmath140 choices for each of the two potentially nonzero entries , hence @xmath141 such matrices in total ( even taking into account the several different choices of which two entries are nonzero ) . in the latter case , there are exactly two corresponding quadruples @xmath78 of integers as in lemma [ baby.enumeration ] . taking into account
that each entry of @xmath42 must be at most @xmath6 in absolute value , we deduce that @xmath142 where @xmath143 is defined as above .
because of the symmetries @xmath144 , we have @xmath145 the only term in the sum where @xmath146 is the term @xmath147 , and for all other terms we can invoke the additional symmetry @xmath148 , valid by switching the roles of @xmath123 and @xmath87 in the definition of @xmath143 .
we obtain @xmath149 where the last step used the fact that @xmath150 . using the characteristic property of the mbius function
, we can write the last expression as @xmath151 as claimed . let @xmath6 and @xmath8 be integers with @xmath152 , and let @xmath153 and @xmath154 be integers with @xmath155 .
then @xmath156 where @xmath157 [ converting from n to cd lemma ] we have @xmath158 since @xmath153 and @xmath154 are positive , we can rewrite this product as @xmath159 where we have used @xmath160 and @xmath161 to slightly simplify the inequalities .
the first factor on the right - hand side of equation is @xmath162 if this expression is positive , and 0 otherwise ; it is thus precisely @xmath163 .
similarly , the second factor on the right - hand side of equation is @xmath164 ( note that this expression is always positive under the hypotheses of the lemma ) , which is simply @xmath165 . multiplying these two factors yields @xmath166 the lemma follows upon noting that both @xmath167 and @xmath168 are @xmath169 by definition , so that the second summand becomes simply @xmath170 , and the @xmath171 term may be subsumed into @xmath170 since @xmath172 .
we have already used the trivial estimate @xmath173 provided @xmath174 .
we will also use , without further comment , the estimates @xmath175 and @xmath176 these estimates ( also valid for @xmath174 ) follow readily from comparison to the integrals @xmath177 and @xmath178 .
most of the technical work in proving theorem [ theorem : particular integer eigenvalue ] lies in establishing an estimate on a sum of the form @xmath179 for a fixed @xmath180 .
the following proposition provides an asymptotic formula for this sum ; we defer the proof until the next section . assuming this proposition , though
, we can complete the proof of theorem [ theorem : particular integer eigenvalue ] , as well as corollary [ corollary : all integer eigenvalues ] .
let @xmath181 and @xmath182 be real numbers , and let @xmath183 and @xmath184 be the functions defined in equation .
then @xmath185 where @xmath20 was defined in equation .
[ key proposition ] the functions @xmath183 and @xmath184 defined in equation are homogeneous of degree @xmath186 in the variables @xmath153 and @xmath154 , so that lemma [ converting from n to cd lemma ] implies @xmath187 inserting this formula into the conclusion of lemma [ initial count and mu insertion ] yields @xmath188 we bound the first error term by summing over @xmath189 to obtain @xmath190 so that we have the estimate @xmath191 we now apply proposition [ key proposition ] to obtain @xmath192 where we have used equation and the fact that @xmath193 and @xmath194 are convergent ( so the partial sums are uniformly bounded ) . note that for any @xmath77 , if one eigenvalue is an integer then they both are ( since the trace of @xmath42 is an integer ) .
thus if we add up the cardinalities of all of the @xmath76 , we get twice the cardinality of @xmath195 , except that matrices with repeated eigenvalues only get counted once . however , the number of such matrices is @xmath112 by lemma [ lemma : double eigenvalue ] .
therefore @xmath196 the sum is a riemann sum of a function of bounded variation , so this becomes @xmath197 the corollary then follows from the straightforward computation of the integral @xmath198 , noting that @xmath199 .
it remains to prove proposition [ key proposition ] . recalling
that the functions @xmath183 and @xmath184 defined in equation are formed by combinations of minima and maxima , we need to separate our arguments into several cases depending on the range of @xmath30 .
the following lemma addresses a sum that occurs in two of these cases ( @xmath200 and @xmath201 ) .
note that because of the presence of terms like @xmath202 in the formula for @xmath20 , we need to exercise some caution near @xmath203 .
let @xmath181 and @xmath204 be real numbers , with @xmath205
. then @xmath206 [ lemma where lower summation limit could be 1 ] suppose first that @xmath207 .
then the sum in question is @xmath208 which establishes the lemma in this case . on the other hand , if @xmath209 then the sum in question is @xmath210 we subtract @xmath211 from the main term and compensate in the error term to obtain @xmath212 since we are working with the assumption that @xmath213 . because the function @xmath214 is increasing on the interval @xmath215 and bounded on the interval @xmath216 $ ] , we have @xmath217 if @xmath218 and @xmath219 if @xmath220 . in either case ,
the last error term can be simplified to @xmath221 , which establishes the lemma in second case .
we consider separately the four cases corresponding to the different parts of the definition of @xmath20 . _
@xmath222 case 1 : @xmath223 .
_ in this case we have @xmath224 and @xmath225 therefore @xmath226 ( the first sum might be empty , but this does not invalidate the argument that follows . ) the first sum is simply @xmath227 by lemma [ lemma where lower summation limit could be 1 ] , the second sum is @xmath228 while the third sum is @xmath229 this case of the proposition then follows from equation , on noting that @xmath230 _ @xmath222 case 2 : @xmath203 .
_ in this case we have @xmath231 therefore @xmath232 as desired . _
@xmath222 case 3 : @xmath233 .
_ in this case we have @xmath234 therefore @xmath235 ( we note that @xmath236 for @xmath30 between 1 and @xmath237 . for very small
@xmath180 we might have @xmath238 , in which case the first sum is empty , but that does not invalidate the argument that follows . ) by lemma [ lemma where lower summation limit could be 1 ] , the first sum is @xmath239 while the second sum has already been evaluated in equation above .
this case of the proposition then follows from equation , on noting that @xmath240 _ @xmath222 case 4 : @xmath241 . _ just as in case 3 , we have @xmath234 however , the inequality @xmath242 automatically implies that @xmath243 when @xmath244 .
therefore @xmath245 ( in this case we will not need to use the more precise lower bound @xmath246 . )
this yields @xmath247 where the error terms have been simplified since @xmath248 is bounded away from 0 .
this ends the proof of proposition [ key proposition ] .
in proving theorem [ theorem : distribution of real eigenvalues ] , it will be convenient to define the odd function @xmath249 whose relevance is demonstrated by the following lemma .
if @xmath74 and @xmath183 are independent random variables uniformly distributed on @xmath2 $ ] , then the product @xmath250 is a random variable whose distribution function is @xmath251 for @xmath252 $ ] .
[ product distribution ] of course , for @xmath253 we have @xmath254 , and likewise @xmath255 for @xmath256 .
note that @xmath257 and @xmath258 are uniformly distributed on @xmath259 $ ] . for @xmath260
, we easily check that @xmath261 .
thus @xmath262 is distributed on @xmath259 $ ] with density @xmath263 , and by symmetry @xmath250 has density @xmath264 on @xmath2 $ ] .
the lemma follows upon computing @xmath265 .
it will also be helpful to define the following functions , which are symmetric in @xmath153 and @xmath154 : @xmath266 to prove theorem [ theorem : distribution of real eigenvalues ] , we first consider the distribution function @xmath267 associated to the density @xmath41 . for a random matrix @xmath42 in @xmath40)}$ ] and a real number @xmath30
, we will derive an expression for the expected number of real eigenvalues of @xmath42 falling below @xmath30 , then differentiate it to obtain @xmath41 . since the set @xmath40)}$ ]
is closed under negation , it is clear that @xmath268 , so it suffices to compute @xmath41 for @xmath269 $ ] .
it turns out that our calculations for @xmath270 will be somewhat simplified by considering @xmath271 rather than @xmath272 .
we have @xmath273 for all @xmath274 , where @xmath275 is defined in equation . [ f_w proposition ]
let us denote the entries of @xmath42 by the random variables @xmath276 , @xmath74 , @xmath183 , @xmath184 , which by assumption are independent and uniformly distributed in @xmath2 $ ] .
let @xmath30 be fixed in the range @xmath277 $ ] , and consider the shifted matrix @xmath278 , which we write as @xmath279 where @xmath74 , @xmath183 are as before , and @xmath280 , @xmath281 range independently and uniformly in @xmath282 $ ] .
clearly the eigenvalues of @xmath42 less than @xmath283 correspond to the negative ( real ) eigenvalues of @xmath284 . by lemma
[ lemma : double eigenvalue ] , we are free to exclude the null set where @xmath284 is singular or has repeated eigenvalues . outside of this null set , @xmath284 has exactly one negative eigenvalue if and only if @xmath285 , by lemma [ discriminant](c ) . likewise by lemma [ discriminant](d )
, @xmath284 has exactly two negative eigenvalues if and only if @xmath286 and @xmath287 and @xmath288 .
we thus have : @xmath289 we may express this probability as the average value @xmath290 where for fixed @xmath153 and @xmath154 , @xmath291 ( here @xmath292 denotes the indicator function of the indicated relation ) . to complete the proof it suffices to show that @xmath293 equals the function @xmath275 defined in equation .
the probabilities appearing in equation are effectively given by lemma [ product distribution ] .
however , there is some case - checking involved in applying this lemma , since the value of , say , @xmath294 depends on whether @xmath295 , @xmath296 , or @xmath295 .
we make some observations to reduce the number of cases we need to examine .
note that @xmath297 is bounded between 0 and 1 for any @xmath298 $ ] , so that @xmath299 always lies in the interval @xmath2 $ ] prescribed by lemma [ product distribution ] . from the identity @xmath300 we see also that @xmath301 .
thus @xmath302 is never lower than @xmath186 , and we need only consider whether @xmath303 ( in which case @xmath304 ) .
we therefore have @xmath305 and @xmath306 inserting these two evaluations into the formula , we obtain @xmath307 it can be verified that this last expression is indeed equal to the right - hand side of the definition of @xmath275 . since @xmath308 , to finish the proof of theorem [ theorem : distribution of real eigenvalues ] it therefore suffices to prove that @xmath309 equals the formula given in equation .
proposition [ f_w proposition ] expresses @xmath272 as an integral , of a function @xmath275 that is independent of @xmath30 , over the square @xmath310 ^ 2 $ ] .
since the region @xmath311 varies continuously with @xmath30 , we can compute the derivative @xmath309 by an appropriate line integral around the boundary of @xmath311 .
indeed , by the fundamental theorem of calculus , we have @xmath312 where we have used the symmetry @xmath313 to reduce the integral to just the top and bottom edges of @xmath311 ( where @xmath314 and @xmath315 , respectively ) . for this range of @xmath30 , the line @xmath318 intersects the bottom edge of @xmath311 at @xmath319 , while the hyperbola @xmath320 intersects the top edge at @xmath321 .
thus by the definition of @xmath275 , equation becomes @xmath322 the following elementary antiderivatives , which are readily obtained by substitution and integration by parts , follow for any fixed nonzero real number @xmath154 from the definitions , , and of @xmath323 , @xmath324 , and @xmath325 : @xmath326 therefore in this case @xmath327 ( after some algebraic simplification ) , which verifies the first case of theorem [ theorem : distribution of real eigenvalues ] .
( note that the integrands really are continuous , despite terms that look like @xmath328 , because the function @xmath323 is continuous at 0 ; hence evaluating the integrals by antiderivatives is valid . ) now , the line @xmath318 does not intersect @xmath311 , while the hyperbola @xmath320 intersects the top edge at @xmath321 .
thus by the definition of @xmath275 and the antiderivative of @xmath325 , equation becomes @xmath329 which verifies the second case of theorem [ theorem : distribution of real eigenvalues ] .
as before , the line @xmath318 does not intersect @xmath311 , while the hyperbola @xmath320 intersects the bottom edge at @xmath330 . thus by the definition of @xmath275 and the antiderivative of @xmath325 , equation becomes @xmath331 which verifies the third case of theorem [ theorem : distribution of real eigenvalues ] .
one could also use the same method to extract the individual distributions of the greater and lesser eigenvalues of @xmath42 : for instance , eliminating the factor of 2 from equation would yield an expression for the distribution of just the lesser eigenvalue of @xmath42 . | random matrices arise in many mathematical contexts , and it is natural to ask about the properties that such matrices satisfy . if we choose a matrix with integer entries at random , for example , what is the probability that it will have a particular integer as an eigenvalue , or an integer eigenvalue at all ?
if we choose a matrix with real entries at random , what is the probability that it will have a real eigenvalue in a particular interval ?
the purpose of this paper is to resolve these questions , once they are made suitably precise , in the setting of @xmath0 matrices . |
the remarkable success of the mean - field bardeen - cooper - schrieffer ( bcs ) theory in describing classic superconductors @xcite may be attributed to the fact that an infinitesimal attraction between electrons suffices to form cooper pairs responsible for superconductivity.@xcite on the other hand , we now have an increasing number of superconductors and fermi superfluids where relevant interactions are stronger beyond the applicability of the mean - field theory .
they include superfluid @xmath2he at high pressure , @xcite heavy - fermion superconductors with competing fluctuations,@xcite high-@xmath3 cuprate superconductors,@xcite and the bcs - bec ( bose - einstein condensation ) crossover in trapped atomic gases.@xcite the present paper is devoted to developing a convenient systematic method to describe those systems based on many - body quantum field theory . to be specific , we will focus on a two - body interaction and extend the luttinger - ward self - consistent perturbation expansion for normal states@xcite to superconductors in such a way that , given a normal - state approximation , all the pair ( i.e. , `` anomalous '' ) processes derivable from it are automatically incorporated . reproducing the hartree - fock theory as the lowest - order approximation , the luttinger - ward formalism enables us to include correlation effects systematically and microscopically up to a desired order .
its key ingredient is a functional @xmath4 $ ] of the single - particle green s function @xmath5 described by closed skeleton feynman diagrams , as properly identified and called `` @xmath6-derivable approximations '' by baym.@xcite indeed , choosing an approximate @xmath6 enables us to calculate the whole thermodynamic series ranging from the thermodynamic potential and single - particle green s function to the two - particle and higher - order green s functions.@xcite to put it another way , @xmath6 uniquely determines how to resolve the bogoliubov - born - green - kirkwood - yvon ( bbgky ) hierarchy.@xcite an additional prominent advantage of the formalism is that it automatically satisfies conservation laws , as shown by baym .
@xcite hence , one can also study non - equilibrium phenomena by only changing the imaginary - time matsubara contour into the real - time schwinger - keldysh contour .
@xcite it should be noted that a formal extension of the luttinger - ward theory to superconductors was performed by de dominicis and martin already in 1964.@xcite however , they did not present any practical method of how to efficiently collect pair processes with anomalous green s functions , whose number grows exponentially in comparison with normal diagrams as we proceed to higher orders .
for example , ( a)-(j ) in fig .
[ fig1 ] exhaust the third - order diagrams for @xmath6 in terms of the symmetrized vertex @xmath7 of abrikosov _
et al_.,@xcite where anomalous diagrams ( c)-(j ) are already four times as large in number as normal diagrams ( a ) and ( b ) . .
a square ( a small circle ) represents the bare vertex @xmath7 ( @xmath8 ) that is symmetric in the space - spin coordinates ( both the space - spin coordinates and particle - hole indices ) .
the number below each of diagrams ( a)-(j ) denotes its relative weight .
[ fig1 ] ] here , we develop a perturbation expansion of gathering all the pair processes systematically and concisely without drawing anomalous diagrams for both the singlet and triplet pairings or their mixtures .
this is made possible by introducing a bare interaction vertex @xmath8 that is further symmetrized in the particle - hole ( i.e. , creation - annihilation or `` charge '' ) indices .
thus , processes ( a)-(j ) in fig .
[ fig1 ] can be represented by a single diagram on the right - hand side with a definite analytic expression , i.e. , eq .
( [ phi_3 ] ) below , which reproduces their weights correctly .
the key functional @xmath6 is defined as a closed skeleton expansion in terms of the matrix green s function @xmath1 in the nambu space , which may be approximated by some selected terms from the whole series in practical calculations .
besides the dyson - gorkov equation for @xmath1 and an expression of the thermodynamic potential , we also derive the bethe - salpeter equation for the two - particle green s function that can be solved based solely on @xmath6 .
the formalism enables us to study not only superconductivity itself but also competing effects of charge , spin , and superconducting fluctuations or orders simultaneously through the matrix structure of @xmath8 .
this last point may be regarded as a definite advantage of the present approach over the one based on the stratonovich - hubbard transformation,@xcite where only a single fluctuation relevant to the auxiliary field may be taken into account after some standard approximation . using the formalism
, we also extend the fluctuation exchange ( flex ) approximation originally developed for normal states@xcite so as to incorporate all the pair processes for the spin - singlet pairing .
the flex approximation has been successful in describing anomalous normal - state transport phenomena of high-@xmath3 superconductors as well as some organic and heavy - fermion superconductors.@xcite its extension to superconductivity as described above , which has not been performed yet and will be called `` flex - s approximation '' below , will be useful for understanding those superconductors quantitatively .
this will be carried out in a transparent and concise manner in terms of @xmath8 in contrast to the original derivation for normal states.@xcite naturally , considerable efforts have been made to go beyond the bcs theory .
eliashberg incorporated the electron - phonon interaction explicitly as the source of the attraction to derive `` strong - coupling '' equations of superconductivity within the second - order perturbation in terms of the renormalized electron green s function and bare phonon propagator.@xcite leggett seriously considered interactions between particles other than the pairing part to develop a theory of superfluid fermi liquids at low temperatures.@xcite anderson and brinkman @xcite presented the idea of `` feedback effects , '' i.e. , a change of the pairing interaction through the superfluid transition , to understand the a phase of superfluid @xmath9he that is realized against the mean - field balian - werthamer theory.@xcite the idea was elaborated by brinkman _ et al_. @xcite and kuroda,@xcite who both incorporated the first few anomalous processes into the normal - state pairing interaction composed of the particle - hole `` paramagnon '' diagrams@xcite and evaluated the extra free energy as a power series in the @xmath10-wave energy gaps .
subsequently , tewordt @xcite refined this approach into a self - consistent @xmath6-derivable approximation , where he included those pair processes that can be obtained from the normal particle - hole diagrams by successively replacing a pair of normal green s functions connecting adjacent vertices by a pair of anomalous green s functions ; relevant diagrams in the third order are ( a ) , ( c ) , ( f ) , and ( j ) of fig .
tewordt and coworkers@xcite later applied the formalism to the two - dimensional hubbard model to clarify quasiparticle and spin excitations of high-@xmath11 superconductors .
the same formalism was used by other groups for the two - dimensional hubbard model@xcite and @xmath12-@xmath10 model.@xcite on the other hand , eagles@xcite and leggett@xcite discussed the bcs - bec crossover of the isotropic spin - singlet pairing at zero temperature by combining the bcs gap equation with the equation for the chemical potential .
the transition temperature of the bcs - bec crossover problem was calculated subsequently by nozires and schmitt - rink@xcite based on the thouless criterion.@xcite haussmann and coworkers@xcite studied this problem with an improved @xmath6-derivable approximation to include interaction effects among unpaired and paired fermions . specifically , their @xmath6 consists of the normal particle - particle ladder diagrams appropriate for low - density systems , plus those anomalous processes obtained from the formers by successively replacing two pairs of normal green s functions connecting adjacent vertices by two pairs of anomalous green s functions ; relevant diagrams in the third order are ( b ) and ( h ) of fig . 1 , disregarding ( e ) and ( g ) which may be important even within the ladder approximation .
we should also mention an alternative approach on the issue where some effective pairing interaction is used to solve the eliashberg equations.@xcite however , there is an ambiguity as to how to construct the effective interaction .
similarly , the flex approximation has sometimes been augmented rather phenomenologically to explain the pseudogap phenomena observed in underdoped cuprate superconductors.@xcite to be specific , the `` flex+t - matrix '' approximation@xcite consists of ( i ) identifying the pairing interaction obtained by the flex approximation as the `` bare '' interaction in the thouless criterion and ( ii ) incorporating the contribution of the two - particle `` superconducting fluctuations '' additionally into the single - particle self - energy .
however , the procedure may contain some double counting of elementary processes and has yet to be examined on a firm microscopic basis .
note in this context that the thouless criterion was originally derived by the particle - particle ladder approximation in the bare perturbation expansion for a two - particle green s function,@xcite reproducing the same @xmath11 equation as the mean - field bcs theory .
thus , one may be convinced that we still do not have practical methods of incorporating all the anomalous processes that are naturally present below @xmath11 .
this paper is organized as follows . in 2 , we develop a basic formalism to calculate the thermodynamic potential , single - particle green s function , and two - particle green s function of superconductivity self - consistently based on a single functional @xmath0 $ ] , whose expression may be obtained by a concise perturbation expansion in terms of @xmath8 .
we carry it out in the coordinate representation so that general inhomogeneous systems may be handled . in 3 , we focus on homogeneous systems and transform the results of 2 into the momentum-``energy '' representation .
closed equations in the flex - s approximation are also derived .
section 4 presents a brief summary to indicate main results .
the whole contents of the present paper may be regarded as the fermi - superfluid counterpart of the self - consistent perturbation expansion recently developed for bec.@xcite we set @xmath13 throughout with @xmath14 the boltzmann constant .
the system we consider consists of identical particles with mass @xmath15 and spin @xmath16 described by the hamiltonian @xmath17 with [ hamil2 ] @xmath18 @xmath19 here @xmath20 with @xmath21 and @xmath22 denoting the space and spin coordinates , respectively,@xcite @xmath23 and @xmath24 are the creation and annihilation operators of the fermion field , respectively,@xcite @xmath25 with @xmath26 the chemical potential , and @xmath27 is the interaction potential . though disregarded here , the effect of a lattice or trap potential may be included easily in @xmath28 .
the formulation in this section will be carried out in the coordinate representation so that general inhomogeneous systems can be handled .
it is also applicable to lattice models such as the hubbard model , for which every integration over @xmath29 above should be replaced by a summation over the lattice sites .
let us introduce the heisenberg representations of the field operators in the matsubara formalism by @xcite @xmath30 where argument @xmath31 in the round brackets is defined by @xmath32 and variable @xmath33 lies in @xmath34 with @xmath35 denoting the temperature .
the operators @xmath36 and @xmath37 are denoted as @xmath38 and @xmath39 by abrikosov _
et al_. @xcite , respectively ; distinguishing them by the subscript @xmath40 ( i.e. , the particle - hole or `` charge '' index ) as above enables us to simplify the notation and formulation considerably , as seen below . using @xmath41 , we define green s functions by @xmath42 where @xmath43 is the `` time''-ordering operator and @xmath44 denotes the grand - canonical average in terms of @xmath45.@xcite the standard normal - state green s function @xmath5 corresponds to @xmath46 . the elements satisfy @xmath47 with the superscript @xmath48 denoting the complex conjugate .
the nambu matrix @xmath49 is written explicitly as @xmath50 it obeys the dyson - gorkov equation @xmath51 \hat{g}(\bar{3},2)=\hat{\sigma}_{0 } \delta(1,2 ) , \label{dg}\ ] ] where @xmath52 is the @xmath53 unit matrix , @xmath54 , and integrations over barred arguments are implied .
the quantity @xmath55 is defined by @xmath56 where @xmath57 is the third pauli matrix .
finally , @xmath58 in eq .
( [ dg ] ) denotes the self - energy matrix @xmath59 it follows from eqs .
( [ g_ij - symm ] ) , ( [ dg ] ) , and ( [ g_0 ^ -1(1,2 ) ] ) that @xmath58 satisfies the same symmetry relations as @xmath1 in eq .
( [ g_ij - symm ] ) .
as shown by de dominicis and martin,@xcite we can express the thermodynamic potential @xmath60 as a functional of @xmath1 , i.e. , @xmath61 $ ] , so as to satisfy @xmath62 let us write @xmath63 $ ] as @xmath64+\phi , \label{omega}\ ] ] where @xmath65 $ ] , @xmath66 $ ] , and the operator @xmath67 for the nambu matrices is defined by @xmath68 with the subscript of @xmath69 denoting an extra infinitesimal positive constant in @xmath33 to place creation operators to the left of annihilation ones for the equal - time average . using eqs .
( [ dg ] ) and ( [ omega ] ) , one may show easily that eq .
( [ domega / dg=0 ] ) is transformed into a relation between @xmath58 and @xmath6 as @xmath70 equation ( [ omega ] ) with eqs .
( [ dg ] ) and ( [ sigma - phi ] ) forms an extension of the normal - state luttinger - ward functional@xcite to superconductors .
note that we have given @xmath71 as a functional of @xmath1 following baym@xcite instead of @xmath58 in the original treatment.@xcite it follows from eqs .
( [ dg ] ) , ( [ omega ] ) , and ( [ sigma - phi ] ) that we can calculate @xmath1 and @xmath71 self - consistently once @xmath66 $ ] is given explicitly .
the particle number @xmath72 is obtained by @xmath73 which may be used to change an independent variable from @xmath26 to @xmath72 by the legendre transformation @xmath74 .
perturbation expansions of @xmath71 and @xmath1 in terms of @xmath75 in eq .
( [ h_int ] ) can be carried out conveniently with the s - matrix@xcite @xmath76 .
\label{s - matrix}\ ] ] here @xmath77 ( @xmath78 ) are the interaction representations of the field operators given explicitly by @xmath79 and @xmath80 , and @xmath7 denotes the symmetrized bare vertex@xcite @xmath81 , \label{gamma^(0)}\end{aligned}\ ] ] satisfying @xmath82 .
the interaction in eq .
( [ s - matrix ] ) can be expressed graphically as fig .
[ fig2 ] .
that is symmetric in a pair of outgoing ( incoming ) lines .
[ fig2],height=37 ] now , the key functional @xmath6 in eq .
( [ omega ] ) is given exactly in terms of eq .
( [ s - matrix ] ) as @xmath83 = -\frac{<{\cal s}>_{0{\rm c}}-1}{\beta}\biggr|_{{\rm skeleton},\hat{g}_{0}\rightarrow \hat{g } } , \label{phi}\ ] ] where subscript @xmath84 denotes the thermodynamic average with @xmath85 and another subscript @xmath86 implies retaining only connected feynman diagrams .
the right - hand side may be expressed graphically by the skeleton diagrams ( i.e. , diagrams without self - energy insertions ) in the perturbation expansion for @xmath71 with @xmath87 replaced by @xmath1.@xcite approximating eq .
( [ phi ] ) by a few terms or some partial series , we can construct a self - consistent approximation , as already noted in the last paragraph of [ subsec : tp ] .
equation ( [ phi ] ) was proved by luttinger and ward for normal states.@xcite however , it also holds true for superconductors by including all the pair processes .
the feynman rules for the perturbation expansion of eq .
( [ phi ] ) with eq .
( [ s - matrix ] ) was given by abrikosov _
et al_. @xcite for normal states . following them , however , we need to calculate numerical factors from a separate expansion in terms of @xmath27
. this may be the reason why the expansion with @xmath7 has not been used widely .
it is shown in appendix[appendix : feynman ] that the rules can be simplified considerably so that one may perform it directly in terms of @xmath7 , even for superconductors .
for example , the first - order contribution to @xmath6 is graphically given by fig .
[ fig3 ] , which can be expressed analytically as @xmath88 , \label{phi_1}\end{aligned}\ ] ] where @xmath5 , @xmath89 , and @xmath90 are defined together with @xmath91 by [ gf - def ] @xmath92 with eq .
( [ gf - def ] ) , we have incorporated the first symmetry of eq .
( [ g_ij - symm ] ) manifestly into @xmath93 for the purpose of deriving @xmath8 below . .
the number below each diagram denotes its relative weight.[fig3],height=64 ] the expansion with @xmath7 is still cumbersome for collecting anomalous diagrams in superconductivity , however , as seen from fig .
hence , we further modify the expression of eq .
( [ s - matrix ] ) so that the asymmetry between the incoming and outgoing lines in fig .
[ fig2 ] is removed . to be specific , we introduce a bare vertex @xmath94 from eq .
( [ phi_1 ] ) by [ hatgamma^(0 ) ] @xmath95 a straightforward calculation using the symmetry of eq .
( [ gamma^(0 ) ] ) shows that results of all the above differentiations may be summarized as @xmath96 which satisfies @xmath97 now , we can express eq .
( [ s - matrix ] ) alternatively as @xmath98 , \label{s - matrix2}\end{aligned}\ ] ] where @xmath99 denotes the normal - ordering operator of placing creation operators to the left of annihilation operators with a sign change per every permutation of a pair of adjacent field operators.@xcite the equivalence between eqs .
( [ s - matrix ] ) and ( [ s - matrix2 ] ) may be checked easily by substituting eq .
( [ hatgamma^(0)-2 ] ) into the latter and writing the resultant expression without the @xmath99 operator .
note that @xmath99 is only relevant to equal - time averages of the first order in the perturbation expansion .
that is symmetric in four external lines .
[ fig4],height=37 ] the interaction in eq .
( [ s - matrix2 ] ) may be expressed graphically as fig .
[ fig4 ] , which is symmetric in the four external lines . using eq .
( [ s - matrix2 ] ) , we can perform the perturbation expansion for @xmath6 ( and also for any other quantities ) concisely in such a way that all the anomalous processes are incorporated automatically .
figure [ fig5 ] enumerates first- to forth - order diagrams for @xmath6 .
thus , each of the first- to third - order contributions is exhausted by a single diagram , showing manifestly the advantage of using eq .
( [ s - matrix2 ] ) . moreover , the symmetry of eq .
( [ hatgamma - symm ] ) brings the simplification that every possible wick decomposition for a distinct diagram yields the same contribution .
hence , we only need to consider a convenient decomposition ( i.e. , connection of vertices ) for each distinct diagram and multiply the result by the number of possible connections , which may be calculated easily based on a combinatorial consideration .
now , the feynman rules for the expansion of eq .
( [ phi ] ) with eq .
( [ s - matrix2 ] ) are summarized as follows . 1 .
draw all the @xmath100th - order closed skeleton diagrams that are topologically distinct .
for each such diagram , associate the factor @xmath101 .
2 . for each small circle , associate @xmath102 .
3 . identify the number @xmath103 of possible connections of vertices for diagram @xmath104 under consideration .
4 . consider a specific connection of vertices for diagram @xmath104 where every primed argument is linked with an unprimed argument , and associate @xmath105 for each line connecting @xmath106 and @xmath107 .
5 . identify the number @xmath108 of permutations to realize the connection . with the choice in ( d ) , we may find it by replacing every vertex as @xmath109 and equating @xmath108 with the number of closed particle loops in the resultant normal diagram with @xmath27.@xcite 6 .
multiply the expression by @xmath110 .
7 . when calculating the two - particle irreducible vertex @xmath111 by eq .
( [ gamma^(ir ) ] ) below , replace every green s function as @xmath112 so as to incorporate the first symmetry of eq . (
[ g_ij - symm ] ) manifestly in @xmath111 .
rule ( g ) is relevant only to two - particle and higher - order green s functions . .
[ fig5],height=52 ] as an example of rule ( c ) above , let us consider diagram ( 3 ) in fig . [ fig5 ] .
its wick decomposition may proceed as follows : start from an arbitrary operator @xmath113 ( labelled @xmath114 operator in @xmath115 ) and find its partner ( labelled @xmath116 operator in @xmath117 ) from @xmath118 possible candidates ; to connect another line between @xmath115 and @xmath117 , pick out a pair of operators @xmath119 and @xmath120 from @xmath121 possible choices ; to connect a pair of lines between @xmath117 and @xmath122 , find the partners of @xmath123 and @xmath124 ( labelled @xmath125 and @xmath126 operators , respectively ) from @xmath127 possible choices in @xmath122 ; finally , select one of the @xmath128 possibilities to connect a pair of lines between ( @xmath129,@xmath130 ) and ( @xmath131,@xmath132 ) .
thus , the number @xmath133 is obtained as @xmath134 now , rules ( a)-(f ) enable us to write down an analytic expression for @xmath135 as @xmath136 with @xmath137 from rule ( e ) for this connection .
an elementary calculation using eq .
( [ hatgamma^(0)-2 ] ) shows that eq .
( [ phi_3 - 0 ] ) correctly reproduces the weights of diagrams ( a)-(j ) in fig .
it is convenient at this stage to introduce matrices @xmath8 and @xmath138 by [ ugamma - uchi ] @xmath139 using them , we can express eq .
( [ phi_3 - 0 ] ) concisely as [ phi_1 - 4 ] @xmath140 where the basis of tr is given by the bracket vectors in eq .
( [ ugamma - uchi ] ) the above consideration may be extended to the other diagrams in fig .
[ fig5 ] , where the numbers of rule ( c ) are identified as @xmath141 , @xmath142 , @xmath143 , and @xmath144 .
we thereby obtain @xmath145 @xmath146 @xmath147 @xmath148 the equivalence between eqs .
( [ phi_1 ] ) and ( [ phi_1 - 2 ] ) may be checked easily by using eqs .
( [ gf - def ] ) , ( [ hatgamma^(0)-2 ] ) , and ( [ ugamma - uchi ] ) .
now , one may be convinced that eq .
( [ s - matrix2 ] ) considerably simplifies the perturbation expansion of superconductivity . as already mentioned , eq .
( [ phi ] ) forms a convenient basis for systematic approximations for superconductivity . to be specific ,
green s function and the self - energy are determined self - consistently by eqs .
( [ dg ] ) and ( [ sigma - phi ] ) based on some approximate @xmath6 .
the lowest - order approximation corresponds to the choice @xmath149 from the series of eq .
( [ phi_1 - 4 ] ) , which is exactly the mean - field bcs theory.@xcite it may be improved by including the next - order term as @xmath150 , where two - body quasiparticle scatterings are included .
hence , adopting this approximation on the schwinger - keldysh contour,@xcite we can describe thermalization of superconductors , for example .
we now introduce the flex - s approximation , which adds the sort of diagrams ( 3 ) and ( 4a ) in fig . 5 up to infinite order besides @xmath93 and @xmath151 .
as seen from fig .
1 , it contains the normal particle - hole and particle - particle diagrams , i.e. , diagrams ( a ) and ( b ) , respectively , where every exchange process is automatically incorporated through @xmath7 , plus all the pair processes derivable from them by successively changing directions of a pair of incoming and outgoing arrows at a vertex , i.e. , diagrams ( c)-(j ) . generalizing the consideration that led to eqs .
( [ phi_3 ] ) and ( [ phi_4a ] ) , we find the @xmath100th - order contribution @xmath152 ( @xmath153 ) as @xmath154 the functional @xmath6 of our flex - s approximation is defined by @xmath155 , i.e. , @xmath156 .
\label{phi_flex}\end{aligned}\ ] ] the corresponding self - energy is obtained by eq .
( [ sigma - phi ] ) with @xmath157 , where the differentiation may also be carried out graphically by removing a single line from the diagrams for @xmath6 in all possible ways . using eq .
( [ hatgamma - symm ] ) , we arrive at the expression @xmath158 g_{\bar{k}\bar{l}}(\bar{3},\bar{4 } ) , \label{sigma_flex}\end{aligned}\ ] ] where @xmath159 is defined by @xmath160 next , we consider the two - particle green s function @xmath161 whose poles define collective modes such as density , spin , and superconducting fluctuations .
functional @xmath0 $ ] also enables us to calculate eq .
( [ calk - def ] ) unambiguously . to see this ,
let us apply an artificial nonlocal potential @xmath162 described by the s - matrix @xmath163\ ! .
\label{cals_ex}\ ] ] green s function in the presence of @xmath164 is defined by@xcite @xmath165 which reduces to eq .
( [ g_ij ] ) as @xmath166 .
now , one may see easily that eq .
( [ calk - def ] ) is obtained from eq .
( [ g_ij - u ] ) by @xmath167 this expression tells us that we only need to know the linear response of @xmath168 to @xmath169 for obtaining the two - particle green s function . to find @xmath170 , we start from the dyson - gorkov equation , which is modified from eq .
( [ dg ] ) into@xcite @xmath171\hat{g}^{\cal u\!}(\bar{2},1')=\hat{\sigma}_{0}\delta(1,1 ' ) , \label{dg - u}\ ] ] with @xmath172 let us change @xmath173 in eq .
( [ dg - u ] ) and subsequently set @xmath174 .
we thereby obtain the first - order equation @xmath175\hat{g}(\bar{2},1 ' ) .
\label{dg - du}\ ] ] using eq .
( [ sigma - phi ] ) for the self - energy , we can express @xmath176 above in terms of @xmath170 as @xmath177 where @xmath178 denotes the `` irreducible '' vertex defined by @xmath179 this differentiation should be performed after the replacement of eq . ( [ tg_ij ] ) in @xmath6 , as already mentioned .
let us substitute eq .
( [ dsigma ] ) into eq .
( [ dg - du ] ) , multiply the resultant equation by @xmath180 from the left , and integrate over @xmath31 . changing the arguments appropriately , we obtain a closed equation for @xmath181 as @xmath182 at this stage , it is convenient to introduce the notations @xmath183 and [ mat ] @xmath184 it follows from eqs .
( [ g_ij - symm ] ) , ( [ calk - def ] ) , ( [ gamma^(ir ) ] ) , and ( [ mat ] ) that matrices @xmath185 ,
@xmath111 , @xmath186 all satisfy @xmath187 the first equality is also obeyed by @xmath138 of eq .
( [ chi^(0 ) ] ) .
using eqs .
( [ chi^(0 ) ] ) , ( [ u ] ) , ( [ vec3 ] ) , and ( [ mat ] ) , we can express eq .
( [ dg - eq ] ) as @xmath188 , or equivalently , @xmath189 finally , we obtain an expression of @xmath190 by eq .
( [ calk - deriv1 ] ) as @xmath191 it may be useful to introduce the full vertex @xmath192 which is nothing but the bethe - salpeter equation@xcite for superconductivity and given graphically by fig .
[ fig6 ] . using @xmath193
, we can express eq .
( [ calk ] ) alternatively as @xmath194 equation ( [ calk ] ) or ( [ calk2 ] ) with eqs .
( [ chi^(0 ) ] ) , ( [ gamma^(ir ) ] ) , and ( [ mat ] ) tells us that we can also calculate the two - particle green s function of eq .
( [ calk - def ] ) once @xmath6 is given explicitly .
( @xmath111 ) , and an internal line represents @xmath1 .
[ fig6],height=37 ] it may be illuminating to write down @xmath111 explicitly in the flex - s approximation .
let us make the replacement of eq .
( [ tg_ij ] ) in eq .
( [ phi_flex ] ) , substitute the resultant @xmath6 into eq .
( [ gamma^(ir ) ] ) , and perform the differentiation . also using the symmetry noted around eq . ( [ gc - symm ] )
, we obtain @xmath195 \nonumber \\ & & \hspace{-10 mm } + \langle 12_{ij}|\underline{\gamma}^{(0)}\underline{\chi}^{(0)}\underline{\gamma}^{(1)}|1'2'_{i'j'}\rangle \nonumber \\ & & \hspace{-10 mm } -\langle 12'_{i,3-j'}|\underline{\gamma}^{(0)}\underline{\chi}^{(0)}\underline{\gamma}^{(1)}|1'2_{i',3-j}\rangle \nonumber \\ & & \hspace{-10 mm } + 2\bigl[\langle 1\bar{3}_{i\bar{k}}|\underline{\gamma}^{(1)}|2\bar{4}'_{{j}\bar{l}'}\rangle \langle2'\bar{4}_{j'\bar{l}}|\underline{\gamma}^{(1)}|1'\bar{3}'_{i'\bar{k}'}\rangle \nonumber \\ & & \hspace{-10 mm } -\langle 1\bar{3}_{i\bar{k}}|\underline{\gamma}^{(1)}|2'\bar{4}'_{3-{j}',\bar{l}'}\rangle \langle2\bar{4}_{3-j,\bar{l}}|\underline{\gamma}^{(1)}|1'\bar{3}'_{i'\bar{k}'}\rangle \bigr ] \nonumber \\ & & \hspace{-10 mm } \times \langle\bar{3}\bar{4}'_{\bar{k}\bar{l}'}|\underline{\chi}^{(0)}|\bar{3}'\bar{4}_{\bar{k}'\bar{l}}\rangle , \end{aligned}\ ] ] which is different from @xmath159 of eq .
( [ gamma^(1)-def ] ) .
thus , a clear distinction among @xmath159 , @xmath111 , and @xmath193 is necessary . before closing this section , we comment on an alternative self - consistent approach adopted by haussmann _
et al_. @xcite of expressing @xmath71 in terms of a renormalized vertex @xmath196 besides @xmath1 .
its theoretical basis was given by de dominicis and martin ; @xcite see also refs . and on this point
. however , the vertices @xmath196 obtained by de dominicis @xcite and used by haussmann _
et al_. @xcite have no internal degrees of freedom describing only density fluctuations , in contrast to our @xmath8 of eq .
( [ hatgamma^(0 ) ] ) .
indeed , constructing @xmath197 $ ] with a matrix structure in @xmath196 seems to be a non - trivial problem .
it will be practically more convenient to use @xmath61 $ ] that can easily be written down solely in terms of the smallest expansion unit @xmath1 .
note in this context that : ( i ) the results obtained by using @xmath197 $ ] should necessarily be reproduced by the approach based on @xmath61 $ ] , as in the case of the particle - particle ladder approximation by haussmann _
et al_. ; @xcite ( ii ) the vertex @xmath196 in @xmath71 is generally different from both the irreducible vertex @xmath198 of eq .
( [ gamma^(ir ) ] ) and the full vertex @xmath199 of eq .
( [ bs ] ) .
we now simplify the formalism of 2 for homogeneous systems .
lattice models may be handled similarly with modifications given in appendix[appendix : hubbard ] .
equations ( [ g_ij ] ) and ( [ hatsigma ] ) and the delta function @xmath200 can be expanded for homogeneous systems as [ gs - exp ] @xmath201 here @xmath202 and @xmath203 with @xmath204 the fermion matsubara frequency ( @xmath205 ) , subscripts @xmath206 denote spin indices , and the summation over @xmath207 is defined by @xmath208 we call the imaginary quantity @xmath209 `` energy '' below . substituting eq .
( [ gs - exp ] ) into eq .
( [ dg ] ) , we can transform the dyson - gorkov equation into the @xmath210 matrix equation @xmath211\hat{g}(\vec{p}\,)=\hat{1 } , \label{dg(p)}\ ] ] where @xmath212 and @xmath213 are defined by @xmath214\ ] ] with @xmath215 and @xmath216 , respectively .
the standard arrangement of the basis vectors for the direct product @xmath217 of the particle - hole ( i.e. , charge ) space c ( @xmath40 ) and spin space s ( @xmath218 ) is given by @xmath219 , @xmath220 , @xmath221 , and @xmath222 .
green s function in this representation reads @xmath223 where each quantity with an underline is a @xmath53 matrix in spin space given as a fourier coefficient of eq .
( [ gf - def ] ) .
it follows from eqs .
( [ g_ij - symm ] ) and ( [ gf - def ] ) that these submatrices satisfy [ gf - symm ] @xmath224 @xmath225 the self - energy matrix may also be expressed as @xmath226 whose submatrices clearly obey the relations of eq .
( [ gf - symm ] ) .
next , we transform @xmath65 $ ] into the momentum - energy representation .
since the expression depends on the approximation we adopt , we specifically consider the flex - s approximation given by eq .
( [ sigma_flex ] ) . to begin with , we expand @xmath227 in eq .
( [ gamma^(0 ) ] ) as @xmath228 where @xmath229 with @xmath230 the boson matsubara frequency ( @xmath231 ) , and the summation over @xmath232 is defined in the same way as eq . ( [ p - sum ] ) with the replacement @xmath233 and @xmath234 .
let us substitute eq .
( [ gamma^(0 ) ] ) into eq . ( [ hatgamma^(0)-2 ] ) , use eqs .
( [ delta - fn ] ) and ( [ v - exp ] ) subsequently , and make some changes of integration variables ; see also appendix[appendix : hubbard ] for details .
we thereby obtain an expansion of eq .
( [ ugamma^(0 ) ] ) as @xmath235 where @xmath236 is defined by @xmath237 and the fourier coefficient is given by @xmath238 .
\label{gamma - p}\end{aligned}\ ] ] note that the arguments @xmath33 , @xmath239 , @xmath240 , and @xmath241 in eq .
( [ hatgamma^(0)-exp ] ) are properly expanded in fermion matsubara frequencies ; this is why we have expressed eq .
( [ v - exp ] ) in terms of boson matsubara frequencies .
using eq .
( [ g_ij - exp ] ) , we can also transform eq .
( [ chi^(0 ) ] ) into @xmath242 with @xmath243 substituting eqs . ( [ gs - exp ] ) , ( [ hatgamma^(0)-exp ] ) , and ( [ hatchi^(0)-exp ] ) into eq .
( [ sigma_flex ] ) , we obtain the self - energy for homogeneous systems in the flex - s approximation as @xmath244 \nonumber \\ & & \hspace{9 mm } \times \underline{\delta\chi}^{(i\alpha , j\beta)}(\vec{p}-\vec{q } ) , \label{sigma_flex(p)}\end{aligned}\ ] ] where @xmath159 is the momentum - energy representation of eq .
( [ gamma^(1)-def ] ) satisfying @xmath245 and @xmath246 is defined by @xmath247 the basis vectors of tr in eq .
( [ sigma_flex(p ) ] ) are given by @xmath248 , which belong to the direct product @xmath249 with dimension @xmath250 .
it is convenient to arrange them as @xmath251 @xmath252 with @xmath253 then , @xmath254 in eq .
( [ chi^(0)(p , q ) ] ) can be represented by a @xmath255 matrix as @xmath256 where each element is a @xmath210 submatrix given in terms of the elements of eq .
( [ hatg(p ) ] ) ; for example , @xmath257 denotes @xmath258 with @xmath259 .
we can also express @xmath260 in eq .
( [ hatgamma^(0)-exp ] ) as @xmath261 where @xmath262 are given by [ gamma - mat^(a , b , c ) ] @xmath263 with @xmath264 denoting the @xmath210 unit matrix and @xmath265 equation ( [ dchi - def ] ) may also be transformed into a matrix representation .
equations ( [ dg(p ) ] ) and ( [ sigma_flex(p ) ] ) form a closed set of self - consistent equations for @xmath266 in the flex - s approximation , which is applicable to an arbitrary pairing symmetry including a mixture of singlet and triplet pairings . by using the representation of eq .
( [ bv2 ] ) , practical calculations may be performed based on the standard matrix algebra plus treatment of integral equations .
some unitary transformations will be helpful for simplifying those calculations .
for example , consider the orthogonal matrix @xmath267 with @xmath268 @xmath269 @xmath270 by using @xmath271 , eq .
( [ gamma - mat ] ) is transformed into the diagonal form @xmath272 where @xmath273 are given by [ tgamma - mat^(a , b , c ) ] @xmath274 with [ v^eo ] @xmath275 the other quantities in eq .
( [ sigma_flex(p ) ] ) may also be expressed in some convenient forms after the transformation .
finally , we briefly consider the two - particle green s function defined by eq .
( [ calk - def ] ) , which obeys eq .
( [ calk2 ] ) .
let us expand the quantities in eq .
( [ mat ] ) as eq .
( [ hatgamma^(0)-exp ] ) , where @xmath276 is obtained as @xmath277 equation ( [ calk2 ] ) is thereby transformed into @xmath278 where @xmath193 is the fourier transform of eq .
( [ bs ] ) obeying @xmath279 equations ( [ calk - p ] ) and ( [ bs - p ] ) correspond to eqs .
( 6 ) and ( 7 ) of leggett for the spin - singlet pairing in his theory of superfluid fermi liquids,@xcite respectively . here
, we can calculate @xmath190 microscopically based on @xmath6 for both the singlet and triplet pairings .
now , we focus on the contact interaction @xmath280 in this case , we can simplify eq .
( [ sigma_flex(p ) ] ) for the self - energy considerably .
indeed , substituting eq .
( [ u - def ] ) into eq .
( [ tgamma - mat ] ) , we observe that only the 1st , 2nd , 3rd , 4th , 10th , and 14th diagonal elements are finite in the diagonal matrix @xmath281 ; the relevant @xmath282 submatrix is given by @xmath283 moreover , absence of momentum dependences in @xmath8 enables us to perform the @xmath284 summations in eqs .
( [ sigma_flex(p ) ] ) and ( [ gamma^(1)(p ) ] ) . using these properties in eq . ( [ sigma_flex(p ) ] )
, we realize that we only need the 1st , 2nd , 3rd , 4th , 10th , and 14th rows and columns of the two matrices [ tchi - tdchi ] @xmath285 @xmath286 these statements on eq .
( [ sigma_flex(p ) ] ) for the contact interaction hold true for an arbitrary pairing symmetry .
we further restrict ourselves to the spin - singlet pairing with no magnetic polarization for simplicity .
the condition implies @xmath287 , @xmath288 , and @xmath289 in eq .
( [ hatg(p ) ] ) .
it may also be possible for this single - component superconductivity to choose the phase of @xmath290 such that its imaginary part originates solely from the @xmath209 dependence as @xmath291 .
these relations are summarized together with eq .
( [ gf - symm ] ) as [ gf - singlet ] @xmath292 similarly , the elements of eq .
( [ hatsigma(p ) ] ) will obey @xmath293 substituting eq .
( [ gf - singlet ] ) into eq .
( [ dg(p ) ] ) , we observe that the @xmath210 matrix equation is divided into a pair of @xmath53 equations that are equivalent to one another .
the one composed of the 1st and 4th rows and columns are given by @xmath294 with @xmath295 and @xmath296 .
the other finite submatrix from the 2nd and 3rd rows and columns are obtained from eq .
( [ dg - singlet ] ) by changing the signs of @xmath297 and @xmath298 simultaneously .
next , we set @xmath299 , @xmath300 in eq .
( [ sigma_flex(p ) ] ) to express @xmath301 and @xmath298 in eq .
( [ dg - singlet ] ) in terms of @xmath302 and @xmath297 .
elementary calculations with eqs .
( [ dchi - def])-([ff - def ] ) , ( [ r - def ] ) , and ( [ gf - singlet ] ) show that the relevant @xmath282 submatrices of eqs .
( [ tchi - tdchi1 ] ) and ( [ tchi - tdchi2 ] ) , which are denoted by @xmath303 and @xmath304 , respectively , are given by [ check - mat ] @xmath305 @xmath306 @xmath307 where the elements of eq .
( [ cchi^(0 ) ] ) are defined by [ chi_gg(q ) ] @xmath308 let us substitute eqs .
( [ cgamma^(0 ) ] ) and ( [ check - mat ] ) into eq . ( [ sigma_flex(p ) ] ) for @xmath299 , @xmath300 . in both cases ,
the resultant tr can be calculated separately for the two submatrices composed of the 2nd , 3rd , and 4th rows and columns and the 1st , 5th , and 6th rows and columns . also using @xmath309 , etc .
, in the second order , we obtain [ sd - flex - s ] @xmath310 @xmath311 here @xmath100 is the particle density @xmath312 with @xmath313 an infinitesimal positive constant,@xcite the functions @xmath314 , @xmath315 , and @xmath316 are defined by [ chi^cs ] @xmath317 @xmath318 @xmath319 with superscripts @xmath320 and @xmath321 denoting `` charge '' and `` spin , '' respectively , and @xmath322 and @xmath323 are given by @xmath324 @xmath325 respectively . setting @xmath326 in eq .
( [ sigma - flex - s ] ) yields the normal - state self - energy @xmath327 . \label{sigma^(n)}\end{aligned}\ ] ] the second term on the right - hand side completely agrees with eq . ( 2.31 ) by bickers and scalapino in the normal flex approximation,@xcite as it should . to find the equation for @xmath3 ,
let us assume the second - order transition and linearize ( i ) the right - hand side of eq .
( [ delta - flex - s ] ) in terms of @xmath89 and ( ii ) @xmath89 itself in eq .
( [ dg - singlet ] ) in terms of @xmath328 as @xmath329 we thereby obtain the linearized self - consistency equation as @xmath330 where @xmath331 is given by @xmath332 \nonumber \\ & & \hspace{0.5 mm } -u^{2}\sum_{\vec{q}}\frac{(g_{\vec{p}+\vec{q}}+g_{-\vec{p}+\vec{q}})g_{\vec{p}^{\,\prime}+\vec{q}}}{\bigl[1+u\chi^{(0)}_{gg}(\vec{q})\bigr]\bigl[1-u\chi^{(0)}_{g\bar{g}}(\vec{q})\bigr]}. \label{tc - flex - s - singlet}\end{aligned}\ ] ] this @xmath331 may be regarded as the pair potential for the spin - singlet pairing without feedback effects.@xcite note in this context that the feedback effects are included naturally in eq .
( [ delta - flex - s ] ) .
equations ( [ dg - singlet ] ) , ( [ sd - flex - s ] ) , and ( [ n ] ) form a closed set of self - consistent equations for the spin - singlet pairing within the flex - s approximation .
the replacement @xmath333 in eq .
( [ sd - flex - s ] ) yields the standard flex approximation for superconductivity with only the particle - hole contribution;@xcite the charge contribution is disregarded further in many cases .
this approximation may be appropriate for the hubbard model near half filling where spin fluctuations are dominant , for example .
on the other hand , neglecting the particle - hole contribution in eq .
( [ sd - flex - s ] ) altogether by @xmath334 and @xmath335 we obtain the particle - particle ladder approximation adopted by haussmann _
et al_. for the bcs - bec crossover problem.@xcite hence , eq . ( [ sd - flex - s ] ) enables us to study the two cases of considerable physical interest on an equal footing , including a crossover from the one limit to the other .
it is also clear that the flex+t - matrix approximation,@xcite which incorporates a contribution of the two - particle green s function into the single - particle self - energy , can not be reproduced as a limit of the flex - s approximation , where both of the single - particle and two - particle green s functions should be studied starting from the single functional of eq .
( [ phi_flex ] ) .
hence , a microscopic foundation of the flex+t - matrix approximation remains to be examined critically by including more terms from eq .
( [ phi ] ) beyond the flex - s approximation .
we have developed a concise self - consistent perturbation expansion for superconductivity in terms of green s function @xmath1 and fully symmetrized bare vertex @xmath8 defined by eqs .
( [ hatg ] ) and ( [ hatgamma^(0 ) ] ) , respectively , so as to incorporate all the pair processes naturally and conveniently . indeed , distinct diagrams for the key functional @xmath0 $ ] of eq .
( [ phi ] ) are exhausted within the fourth order by the five graphs of fig .
corresponding analytic expressions can be written down immediately as eq .
( [ phi_1 - 4 ] ) based on the feynman rules ( a)-(f ) given around eq .
( [ tg_ij ] ) . using an approximate @xmath6
thereby constructed , we can calculate @xmath1 self - consistently by eqs .
( [ dg ] ) and ( [ sigma - phi ] ) , the thermodynamic potential subsequently by eq .
( [ omega ] ) , and also the two - particle green s function of eq .
( [ calk - def ] ) by eq .
( [ calk ] ) with eqs .
( [ chi^(0 ) ] ) , ( [ gamma^(ir ) ] ) , and ( [ mat ] ) . moreover , entropy may be written down explicitly once @xmath1 and @xmath58 are known.@xcite thus , a single procedure of choosing an approximate @xmath6 enables us to calculate the whole thermodynamic hierarchy , and the results may be improved by starting from a better @xmath6 .
various fluctuations and competing orders may be described simultaneously through the matrix structure of @xmath8 .
since the conservation laws are obeyed automatically , the basic formalism of 2 in the coordinate representation can be applied to study non - equilibrium phenomena by changing the matsubara contour into the schwinger - keldysh contour.@xcite the idea of symmetrizing the bare vertex in terms of the particle - hole indices will also be useful for the electron - phonon hamiltonian to simplify its perturbation expansion .
this formalism has been transformed for homogeneous systems into the momentum - energy representation in 3.1 .
we have subsequently derived a closed set of self - consistent equations for the contact potential in the spin - singlet flex - s approximation as eqs .
( [ dg - singlet ] ) , ( [ sd - flex - s ] ) , and ( [ n ] ) , which adds all the pair processes characteristic of superconductivity to the normal flex approximation.@xcite as shown in the paragraph after eq .
( [ tc - flex - s - singlet ] ) , it contains extra terms besides those in the standard flex approximation for superconductivity , and also embraces the particle - particle ladder approximation by haussmann _
et al_. @xcite for the bcs - bec crossover problem .
thus , our flex - s approximation incorporates the particle - hole and particle - particle diagrams on an equal footing with their mixing effects , thereby enabling us to study the bcs - bec problem from low- to high - density regions continuously .
this work is supported by a grant - in - aid for scientific research from the ministry of education , culture , sports , science and technology of japan .
the feynman rules for the normal - state expansion of eq .
( [ phi ] ) with eq .
( [ s - matrix ] ) are summarized as follows . 1 .
draw all the @xmath100th - order closed skeleton diagrams that are topologically distinct .
for each such diagram , associate the factor @xmath336 .
2 . for each vertex , associate @xmath337 .
3 . identify the number @xmath103 of possible connections of vertices for diagram @xmath104 under consideration .
4 . consider a specific connection of vertices for diagram @xmath104 and associate @xmath338 for each line arriving at @xmath339 from @xmath340 .
identify the number @xmath108 of permutations to realize the connection .
we may find it by replacing @xmath341 and equating @xmath108 with the number of closed particle loops in the resultant diagram with @xmath27.@xcite 6 .
multiply the expression by @xmath110 .
these rules can also be used for superconductors , where the number @xmath108 for an `` anomalous '' diagram may only be found by a direct counting of permutations .
a key quantity in the rules is the numerical weight factor @xmath103 . since the two operators @xmath342 in eq .
( [ s - matrix ] ) with the same subscript @xmath343 ( @xmath344 ) are equivalent , the factor may be obtained elementarily by a combinatorial consideration .
for example , the factors for fig .
[ fig3](a ) and ( b ) are trivially identified as @xmath345 and @xmath346 , respectively . together with @xmath347 and @xmath348 in terms of eq .
( [ gf - def ] ) , we arrive at eq .
( [ phi_1 ] ) for fig .
[ fig3 ] following the above rules .
similarly , the weight factors for fig . [ fig1 ] ( a)-(j ) are found as @xmath349 , @xmath350 , @xmath351 , @xmath352 , @xmath353 , @xmath354 , @xmath355 , @xmath356 , @xmath357 , and @xmath358 , each of which corresponds to the number of possible connections of three vertices for realizing the relevant topological structure .
we thereby obtain the relative weights of fig .
here , we discuss the momentum - energy representation of homogeneous lattice models .
first of all , momentum integrations in eqs . ( [ gs - exp ] ) and ( [ v - exp ] )
should be restricted to the first brillouin zone .
this fact may be incorporated manifestly in these formulas by introducing the function @xmath359 on their right - hand sides . with these trivial modifications , we consider the basic two - particle quantities in eq .
( [ ugamma - uchi ] ) . substituting eq .
( [ gamma^(0 ) ] ) into eq .
( [ hatgamma^(0)-2 ] ) , we can write the matrix element of eq .
( [ ugamma^(0 ) ] ) explicitly in terms of @xmath27 as @xmath360 .
\label{ugamma^(0)-exp - a}\end{aligned}\ ] ] let us substitute eqs .
( [ delta - fn ] ) and ( [ v - exp ] ) with eq .
( [ theta ] ) into eq .
( [ ugamma^(0)-exp - a ] ) and transform the resultant expression into the form of eq .
( [ hatgamma^(0)-exp ] ) by changing integration variables .
we thereby obtain the fourier coefficient in eq .
( [ hatgamma^(0)-exp ] ) as @xmath361 .
\label{gamma - p - a}\end{aligned}\ ] ] note that @xmath362 and @xmath363 above may lie outside the first brillouin zone to give a finite value for the matrix element .
similarly , the fourier coefficient of eq .
( [ hatchi^(0)-exp ] ) for lattice models is obtained as @xmath364 by incorporating @xmath365 functions accordingly in other formulas , the whole results of 3 also become applicable to lattice models . | we develop a concise self - consistent perturbation expansion for superconductivity where all the pair processes are naturally incorporated without drawing `` anomalous '' feynman diagrams .
this simplification results from introducing an interaction vertex that is symmetric in the particle - hole indices besides the ordinary space - spin coordinates .
the formalism automatically satisfies conservation laws , includes the luttinger - ward theory as the normal - state limit , and reproduces the bardeen - cooper - schrieffer theory as the lowest - order approximation .
it enables us to study the thermodynamic , single - particle , two - particle , and dynamical properties of superconductors with competing fluctuations based on a single functional @xmath0 $ ] of green s function @xmath1 in the nambu space .
specifically , we derive closed equations in the flex - s approximation , i.e. , the fluctuation exchange approximation for superconductivity with all the pair processes , which contains extra terms besides those in the standard flex approximation . |
in the realm of electronic circuits , the key role of linear resistors , capacitors , and inductances is clear .
but the electronic circuits have grown to maturity since the introduction of active and _ nonlinear _ elements . indeed _ rectifiers _ , voltage limiters , bistable multivibrators , frequency mixers , etc . , are not realizable without nonlinear components such as _ diodes _ and transistors @xcite .
microelectromechanical systems ( mems ) technology has its roots in microelectronics technology , thus naturally much attention has been paid to the nonlinear behavior of mems for use in signal processing , actuation , and sensing applications @xcite . with the ongoing trend of miniaturizing devices to the submicron scale , the gap between surfaces may about @xmath0 . in this regime ,
the casimir force @xcite can cause tiny elements to stick together and cause devices to malfunction @xcite .
tiny elements are also susceptible to wear @xcite , thus the operation lifetime of miniaturized machines is a concern . to remedy these problems , it has been noticed that the lateral casimir force @xcite can intermesh noncontact parts of nanomachines @xcite . in view of designing useful nanoscale mechanical devices ,
the dependence of casimir force between bodies on their material and geometrical properties is still a subject of intense investigation @xcite .
the simplest noncontact nanomachine is composed of one sinusoidally corrugated plate ( rack ) and one sinusoidally corrugated cylinder ( pinion ) subject to an external load , see fig .
the pinion experiences the lateral casimir force @xmath1 $ ] , where @xmath2 , @xmath3 , and @xmath4 denote the corrugation wavelength , pinion displacement , and rack displacement , respectively .
clearly , the output signal @xmath5 is a nonlinear function of the input signal @xmath6 . in other words , the casimir rack and
pinion is a nonlinear dynamical system .
inspired by the electronic rectifier based on the diode , a _ mechanical rectifier _
based on the casimir rack and pinion , has been proposed .
it has been shown that the pinion rotates with a nonzero average velocity and works against an external load , whether the rack position versus time is a uniform @xcite , sinusoidal @xcite , or periodic triangular signal @xcite .
the casimir rack and pinion is not a perfect mechanical rectifier .
firstly , it is not guaranteed that an arbitrary input signal can be rectified .
in other words , an arbitrary motion of the rack does not necessarily leads to an _ upward _ motion of the load .
secondly , the device performance may be degraded due to the thermal noise .
note that the pinion is mounted on an axle which is surrounded by a fluid ( air ) .
the dynamics of such a small pinion with negligible inertia is deeply influenced by the continuous bombardment of fluid molecules . according to the classic work of langevin ,
the total torque on axle can be written as sum of a frictional torque and a random torque .
the fluctuation - dissipation theorem expresses the noise strength in terms of the rotational friction coefficient and the temperature of the system . , corrugation amplitudes @xmath7 and @xmath8 , mean separation @xmath9 , radii @xmath10 and @xmath11 , external load @xmath12 , and heat bath temperature @xmath13 , characterize the system . ] here we study the effect of thermal noise on the strongly damped rack and pinion device .
we assume that the rack position versus time is a periodic multiharmonic signal @xmath14 whose time average is zero . indeed by a proper choice of the number of harmonics @xmath15 , parameters @xmath16 and @xmath17 , any signal of period @xmath18 can be well approximated by fourier series ( [ fo ] ) .
we show that in a large domain of parameter space and at room temperature , the rectifier works well .
we observe an interesting phenomenon : the highly damped rack and pinion does not rectify the celebrated signal @xmath19 , but many other multiharmonic signals . indeed the device may not operate for a certain rack motion whose highest frequency is @xmath20 , but upon adding terms with a _ higher _ harmonic @xmath21 to that motion , the device lifts _ up _ the load . the practical message is clear : there is no advantage in reducing the number of harmonics of the input signal .
the casimir rack and pinion shown schematically in fig . [ fig1 ] , can be characterized with the corrugation amplitudes @xmath8 and @xmath7 , corrugation wavelength @xmath2 , mean separation @xmath9 , length of the cylinder @xmath22 , radius of the cylinder @xmath10 , radius @xmath11 , and external load @xmath12 .
a heat bath at temperature @xmath13 is in contact with the system .
the lateral casimir force exerted on the pinion reads @xmath1 $ ] , where @xmath3 and @xmath4 are the lateral displacement of the pinion and the rack , respectively .
the amplitude of the lateral casimir force @xmath23 depends on the parameters @xmath9 , @xmath2 , @xmath8 , @xmath7 , @xmath10 , @xmath22 , and @xmath13 @xcite . to describe the dynamics of the heavily damped pinion , we rely on the following langevin equation @xmath24- r w + \eta(t ) = 0 \label{asli}.\ ] ] here
@xmath25 quantifies the rotational friction . @xmath26 and @xmath27 are the casimir torque and the external torque , respectively .
the random torque @xmath28 is a stationary gaussian white noise with zero mean and the correlation @xmath29 . according to the fluctuation - dissipation theorem @xmath30 , where @xmath31 is the boltzmann s constant .
we choose @xmath32 , @xmath33 , and @xmath34 as the units of length , time , and velocity , respectively .
we let starred quantities to denote dimensionless variables , e.g. @xmath35 and @xmath36 .
the langevin equation ( [ asli ] ) thus can be written as @xmath37 where @xmath38 and @xmath39 @xmath40 is a gaussian noise with zero mean and the correlation @xmath41 . here
the dimensionless parameter @xmath42 compares the noise strength and the casimir grip .
the dimensionless parameter @xmath43 compares the external torque and the casimir torque .
we have assumed that the rack undergoes periodic motion @xmath44 .
it immediately follows that @xmath45 is a space and time periodic function @xmath46 where @xmath47 .
fourier expansion of @xmath48 reads @xmath49
we access the probability distribution @xmath50 via the fokker - planck equation @xmath51 as stated before , @xmath52 is invariant under the operations @xmath53 and @xmath54 . thus we seek for a steady state probability distribution @xmath50 with the same symmetries . following denisov , hnggi and mateos @xcite , we use the truncated fourier expansion @xmath55 to approximate @xmath50
. the precision of this approximation can be controlled by the parameters @xmath56 and @xmath57 .
the probability distribution @xmath50 is normalized , i.e. @xmath58 .
this implies that @xmath59 and @xmath60 if @xmath61 .
the rest of coefficients can be found by inserting eqs .
( [ expan1 ] ) and ( [ expan2 ] ) into ( [ fok ] ) .
this yields the following set of _ linear _ algebraic equations @xcite @xmath62 for @xmath63 coefficients @xmath64 .
it is advantageous to introduce the one - to - one index transformation @xmath65 and then utilizing standard numerical techniques to find single index variables @xmath66 .
we frequently encounter band diagonal set of linear equations , where nonzero elements of the @xmath67 matrix are along a few diagonal lines adjacent to the main diagonal . to access the solution more rapidly , we use algorithms optimized for band diagonal equations @xcite .
+ 0 cm + 0 cm the average of the random torque @xmath40 is zero , thus from eq .
( [ moaser ] ) we find that @xmath68 .
powered by the probability distribution @xmath50 , we find @xmath69 recalling that the average rack velocity is zero , we find the average pinion velocity @xmath70 in terms of the easily accessible fourier coefficients @xmath71
. this key result allows us to fully investigate the rectifier performance .
inspired by the experimental setup of chen _ et al . _
@xcite , we assume @xmath72 , @xmath73 , and @xmath74 . at @xmath75 , we estimate @xmath76 for @xmath77 and @xmath78 for @xmath79 @xcite . for these values of @xmath2 , @xmath10 , and @xmath80 , the amplitude of the lateral casimir force
does not change significantly as temperature raises to @xmath81 @xcite . typically , the load @xmath82 and @xmath83 , thus @xmath84 .
the rotational friction coefficient @xmath25 is about @xmath85 @xcite .
thus for @xmath76 and @xmath81 , we find @xmath86 and @xmath87 . considering the signal @xmath6
, we assume @xmath88 and @xmath89 , thus @xmath90 and @xmath91 .
now we consider the biharmonic signal @xmath92 the dimensionless amplitudes @xmath93 , @xmath94 , dimensionless frequency @xmath95 , and phase @xmath96 characterize this signal . as a concrete example
, we consider a biharmonic signal with @xmath97 , @xmath98 , @xmath99 , and @xmath100 , see fig .
[ figbi1](a ) . in our following study of the average pinion velocity
, we change one of the parameters and keep other parameters fixed .
figure [ figbi1](b ) demonstrates @xmath101 as a function of noise strength and for various loads .
@xmath102 monotonically decreases as the noise strength increases .
remarkably , even for the load @xmath103 , the average velocity is positive for noise strengths @xmath104 .
hereafter we assume @xmath105 .
figure [ figbi1](c ) shows @xmath102 as a function of @xmath106 .
remarkably , @xmath102 is zero for frequencies smaller than a threshold @xmath107 . here , almost independent of the load , @xmath108 .
@xmath102 is monotonically decreasing when @xmath109 .
for example , @xmath110 when @xmath111 , but shifts to @xmath112 when @xmath103 . for @xmath113 and loads @xmath114 , @xmath102 exhibits two maxima and one minima .
thus to maximize the average pinion velocity , the frequency must be deliberately tuned .
figure [ figbi1](d ) shows that @xmath102 is a nonmonotonic function of @xmath115 .
there exists a window of @xmath115 to ensure that the device lifts up the load .
moreover , @xmath115 can be tuned to maximize @xmath102 . in the case of @xmath103 , we find @xmath102 reaches its maximum @xmath116 at @xmath117 .
@xmath102 is positive in the window @xmath118 .
figure [ figbi1](e ) depicting @xmath102 as a function of @xmath119 , conveys similar messages .
figure [ figbi1](f ) shows the average pinion velocity as a function of the phase .
we find @xmath120 , _
i.e. _ , @xmath102 is symmetric with respect to the line @xmath121 .
moreover , @xmath102 may be positive in _ disconnected _ windows of @xmath122 . for @xmath103 , these windows are @xmath123 , @xmath124 , and @xmath125 radians .
one intuitively expects the domain of parameters ensuring @xmath126 to shrink , as the load increases .
figures [ figbi1](b)-(f ) all confirm this expectation . in the following , we present a few sections of @xmath127 parameter space where the noisy rack and pinion derived by a biharmonic signal , properly lifts upward the external load . except the parameter being varied , @xmath97 , @xmath98 , @xmath99 , and @xmath128
are assumed .
we compare the device performance at no noise @xmath129 , and at room temperature noise @xmath130 .
figures [ figbi2](a ) and ( b ) demonstrate domains of positive pinion velocity in the @xmath131 space for @xmath129 and @xmath130 , respectively . at first look , the noise degrades the performance : for example , at @xmath132 , the maximum load decreases from @xmath133 to @xmath134 as the temperature increases . but quite remarkably , the noise may facilitate the device operation : for @xmath129 , a gap @xmath135 and a gap around @xmath136 indicate that the device does not lift up the load .
on increasing the temperature , the first gap shrinks to @xmath137 and the second gap completely disappears . figures [ figbi2](c ) and ( d ) demonstrate domains of positive pinion velocity in the @xmath138 space for @xmath129 and @xmath130 , respectively .
focusing on the intervals @xmath139 , @xmath140 and @xmath141 , again we find that the noise may facilitate the device operation .
this conclusion is also supported by figs .
[ figbi2](e ) and ( f ) .
+ 0 cm + 0 cm
as an example of multi - harmonic driving signal , we consider @xmath142 where @xmath143 , @xmath144 , @xmath145 , @xmath146 , @xmath147 , @xmath148 , and @xmath149 , see fig .
[ fighexa](a ) .
we assume that the noise strength is @xmath150 .
to study the average pinion velocity , we change one of the above parameters and keep others fixed . figures
[ fighexa](b)-(g ) demonstrate @xmath101 for various loads , as a function of @xmath151 , @xmath106 , @xmath115 , @xmath119 , @xmath152 , and @xmath153 , respectively .
figures [ fighexa](b)-(g ) clearly show the possibility of choosing the parameters to guarantee @xmath154 . however , this domain of parameters shrinks , as the load increases .
figure [ fighexa](b ) shows that for @xmath155 and @xmath156 , indeed @xmath101 first increases , reaches its maximum , and then monotonically decreases as the noise strength @xmath151 increases .
this reminds us the stochastic resonance effect @xcite .
figure [ fighexa](c ) shows that @xmath102 is zero for frequencies smaller than a threshold @xmath107 , cf .
[ figbi1](c ) .
we find @xmath157 when @xmath158 . for loads
@xmath158 , @xmath159 and @xmath160 , the average pinion velocity becomes practically positive when @xmath106 becomes larger than @xmath161 , @xmath162 , and @xmath163 , respectively .
@xmath102 is monotonically decreasing when @xmath109 , cf .
[ figbi1](c ) .
for example , @xmath164 when @xmath111 , but shifts to @xmath165 when @xmath103 .
the behavior of @xmath102 for @xmath113 again warns that the frequency must be deliberately tuned to maximize the average pinion velocity .
figure [ fighexa](c)-(g ) show that @xmath102 may be positive in _ disconnected _ windows of parameters , cf .
[ figbi1](d ) and [ figbi1](f ) .
figures [ fighexa](h ) and [ fighexa](i ) demonstrate domains of positive pinion velocity in the @xmath131 space for @xmath129 and @xmath130 , respectively . the noise may degrade the machine performance : for example , at @xmath166 , the maximum load decreases from @xmath167 to @xmath168 as the temperature increases .
but the noise may facilitate the device operation : for @xmath129 , two gaps @xmath169 and @xmath170 indicate that the device does not lift up the load . on increasing the temperature , the first gap shrinks to @xmath171 and
the second gap completely disappears , cf .
[ figbi2](a)-(b ) .
+ 0 cm + 0 cm as another multi - harmonic signal , we consider @xmath172 which approximates the periodic triangular signal characterized by the durations @xmath173 and @xmath174 , and amplitude @xmath175 , see fig . [ triangu](a ) . @xmath176 and @xmath15 are the fundamental frequency and the number of harmonics , respectively . as a concrete example
, we assume @xmath177 , @xmath149 , @xmath178 , @xmath179 , and @xmath150 .
similar to our previous studies of @xmath180 , we change one of the above parameters and keep others fixed . figures
[ triangu](b)-(e ) demonstrate @xmath101 for various loads , as a function of @xmath151 , @xmath106 , @xmath181 , and @xmath182 , respectively .
we observe that the signal parameters @xmath106 , @xmath181 , and even @xmath182 must be tuned to maximize the average pinion velocity . here
@xmath102 can be as large as @xmath183 , note that in all previous examples @xmath102 is below @xmath184 .
figures [ triangu](b)-(e ) again show the possibility of choosing the parameters to guarantee @xmath154 . as expected , this domain of parameters shrinks , as the external load increases .
figure [ triangu](c ) shows that @xmath102 is monotonically decreasing when @xmath109 . as before , @xmath185 depends on the external load . here
@xmath186 is quite small , cf .
[ figbi1](c ) and [ fighexa](c ) .
figure [ triangu](f ) demonstrates domain of positive pinion velocity in the @xmath131 space for room temperature noise @xmath130 .
note that here loads as large as @xmath187 can be lifted up , while in previous examples admissible loads are below @xmath188 .
we map the langevin equation ( [ asli ] ) describing the pinion dynamics into the eq .
( [ moaser ] ) which is well known in the realm of brownian motors @xcite .
however , one must note that ( i ) the casimir device is intended to rectify a _
deterministic _ periodic input motion ( rack motion ) .
( ii ) only _ positive _ average pinion velocities are of interest , since the device should lift up the load .
this restricts the allowed set of parameters @xmath189 characterizing the system .
one can use the stochastic runge - kutta algorithm of the second order @xcite or similar methods to solve the langevin equation ( [ moaser ] ) .
this is straightforward , but requires a large ensemble of realizations to obtain a reliable average .
we follow denisov , hnggi and mateos @xcite to map the original problem onto a set of linear algebraic equations .
this allows us to explore a huge parameter space in a reasonable computational time ( a few days ) .
input signals that lead to @xmath190 are of no interest .
it is known that @xmath191 @xcite if @xmath192 under one of the transformations @xmath193 specified by the parameters @xmath194 and @xmath195 . according to eq .
( [ myg ] ) , both symmetries @xmath196 and @xmath197 are absent if @xmath198 . for @xmath199 and @xmath19 , the symmetry @xmath196 exists and thus @xmath191 .
now it is not surprising that the device can not lift up a load upon monochromatic driving of the rack . for @xmath200 and biharmonic
driving ( [ twotwo ] ) , the symmetry @xmath197 exists if @xmath201 .
figure [ figbi1](f ) shows that at @xmath202 indeed @xmath203 as @xmath204 . in the case of multiharmonic signals ( [ tritri ] ) and ( [ dah ] ) , both symmetries @xmath196 and @xmath197
are absent even if @xmath200 . for similar studies of brownian motors driven by biharmonic signals ,
see refs .
@xcite .
we observe that the noise may facilitate the device operation .
figures [ figbi2](a)-(f ) and [ fighexa](h)-(i ) show that the gaps in the @xmath205 domain of parameters , close as the temperature increases .
this suggests to search for the fingerprints of the stochastic resonance effect in the gaps of @xmath205 domain of parameters . for the biharmonic
driving , @xmath206 are in the gap , see fig [ figbi2](c ) .
figure [ stores](a ) clearly supports the occurrence of stochastic resonance for these parameters .
figure [ stores](b ) shows the occurrence of stochastic resonance for the polyharmonic signal ( [ tritri ] ) .
here @xmath207 , and dashed and solid lines correspond to @xmath208 and @xmath209 , respectively .
note that @xmath210 for @xmath211 . in units of @xmath212 versus @xmath213 for ( a ) the biharmonic driving ( [ twotwo ] ) with parameters @xmath99 , @xmath214 , @xmath215 , and @xmath216 , ( b ) the polyharmonic signal ( [ tritri ] ) with parameters @xmath217 , @xmath218 , @xmath219 , @xmath220 , @xmath221 , and @xmath222 . ]
our results show that @xmath223 is achievable , see figs .
[ figbi1 ] , [ fighexa ] and [ triangu ] . reminding the typical pinion radius @xmath224 and skipping velocity @xmath225 for the gap @xmath77
, this means that the pinion rotates with an average angular velocity @xmath226 .
quite importantly , the amplitude of the lateral casimir force rapidly grows as the gap @xmath9 decreases . for @xmath79 then @xmath227 and @xmath228
this is a considerable average angular velocity . according to eq .
( [ myds ] ) , the effective noise strength @xmath229 rapidly decreases as the gap size @xmath9 decreases .
assuming @xmath77 , we estimated @xmath230 at room temperature . now reducing the gap size by only a factor of two yields @xmath231 .
thus the noise influence on the the device operation can be controlled and optimized . in summary ,
the casimir rack and pinion acts as a nano - scale mechanical rectifier : even in the presence of the thermal noise , the device can lift up a load when the rack undergoes a periodic multiharmonic motion .
the geometrical parameters , especially the gap @xmath9 , and the input signal parameters , can be tuned to optimize the device performance .
we thank financial support from the `` center of excellence on the structure and physical properties of matter '' of the university of tehran .
we also thank r. sepehrinia and a. a. saberi for their enlightening comments . | we study noncontact rack and pinion composed of a corrugated plate and a corrugated cylinder intermeshed via the lateral casimir force .
we assume that the rack position versus time is a periodic multi - harmonic signal .
we show that in a large domain of parameter space and at room temperature , the device acts as a mechanical rectifier : the pinion rotates with a nonzero average velocity and lifts up an external load .
the thermal noise may even facilitate the device operation . |
over the last decade , the dynamics of ferromagnetic systems below their critical temperatures in a periodically oscillating magnetic field have been studied both theoretically @xcite and experimentally @xcite .
the systems exhibit two qualitatively different behaviors referred to as _ symmetry - restoring oscillation _ ( sro ) and _ symmetry - breaking oscillation _ ( sbo ) , depending on the frequency @xmath0 and the amplitude @xmath1 of the applied magnetic field .
it has been established that there exists a sharp transition line between sro and sbo on the ( @xmath0 , @xmath1 ) plane , which is called the _ dynamical phase transition _ ( dpt ) .
the dpt was first observed numerically in the deterministic mean - field system for a ferromagnet in a periodically oscillating field @xcite , and has subsequently been studied in numerous monte carlo simulations of the kinetic ising system below critical temperature @xcite .
it has also been observed experimentally in an ultra - thin co film on cu(100 ) @xcite .
recently , we investigated the dpt by introducing the model equation @xmath2 @xcite .
this equation is a simplified model for the ising spin system at the temperature @xmath3 below its critical value @xmath4 in an external periodic magnetic field . by appropriately scaling the magnetization @xmath5 , time @xmath6 , and the applied field , this equation is written as @xmath7 the sbo and sro
are observed in eq .
( [ eq : tdgl - scale - external - field ] ) and the transition line between them on the ( @xmath0 , @xmath1 ) plane is determined analytically @xcite .
it is quite interesting to ask whether dpt is observed under another kind of applied field , especially random field with bounded amplitude .
the fundamental aim of the present paper is to study the dynamics of @xmath8 with a dichotomous markov noise ( dmn ) @xmath9 instead of periodically oscillating external field @xmath10 ( see , e.g. , @xcite ) .
the equation of motion @xmath11 with a nonlinear function @xmath12 and the dmn @xmath9 has been extensively studied by many authors @xcite .
it is well known that the master equation for the system can be derived , and then transition phenomena of stationary probability densities concerning the intensity of @xmath9 , for example , are studied , which are referred to as the noise - induced phase transition @xcite .
the asymptotic drift velocity @xmath13 in the case of @xmath12 being periodic functions are also discussed as a specific dynamic property @xcite . furthermore , the mean first - passage time ( mfpt ) and transition rates are investigated as another important dynamic property when @xmath12 is the force associated with the bistable potential given by eq .
( [ eq : general - external - field ] ) @xcite . for a review of works on dmn system ,
see bena @xcite .
the fundamental aim of the present paper is to propose a phenomenological approach to the critical dynamics near the transition point between the symmetry - restoring motion ( srm ) and the symmetry - breaking motion ( sbm ) observed in eq .
( [ eq : general - external - field ] ) with @xmath14 ( sec .
[ sec : model - equation ] ) and to investigate the distribution of passage times through channels , switching times between two different motions , and the power spectrum of the order parameter evolution .
the present paper is constructed as follows . in sec .
[ sec : model - equation ] , we discuss the dynamics of symmetry - breaking motion and symmetry - restoring motion of the model equation ( [ eq : general - external - field ] ) . in sec .
[ sec : average - passage - time ] , the jumping process of the magnetization through channels , which are defined in the text , is investigated and the mfpt is obtained . in sec . [
sec : probabilistic - analysis ] , a phenomenological approach simplifying the dynamics of passing through a channel in the srm phase is introduced and three statistical characteristics are analytically developed .
the results are compared with numerical simulations .
concluding remarks are given in sec .
[ sec : concluding ] .
we consider the equation of motion driven by the external field @xmath9 , @xmath15 where @xmath9 is a symmetric dmn with taking the values @xmath16 . here
the probability @xmath17 that @xmath9 continues to take the identical value @xmath18 or @xmath19 longer than time @xmath20 is given by @xmath21 this implies that the correlation time of @xmath9 is equal to @xmath22 . throughout this paper , numerical integrations of eq .
( [ eq : tdgl - external - field ] ) are carried out by using the euler difference scheme with the time increment @xmath23 . without dmn
, @xmath8 eventually approaches either of the stationary fixed points @xmath24 , one of which is achieved according to the initial condition @xmath25 as shown in fig .
[ fig : f_x ] .
the function @xmath14 is shown by solid lines .
( a ) real roots @xmath26 , ( @xmath27 , 2 , 3 ) for @xmath28 and ( b ) real roots @xmath29 for @xmath30 of the algebraic equation @xmath31 . the definition of @xmath32 is graphically represented .
two bold lines drawn in ( b ) indicate the channel regions defined in sec .
[ sec : average - passage - time ] . ] in the presence of dmn , if @xmath28 , @xmath32 being defined by @xmath33 then @xmath34 has three real roots @xmath35 , ( @xmath27 , 2 , and 3 ) . each value of @xmath36
is graphically shown in fig .
[ fig : f_x](a ) . on the other hand , if @xmath30 , then @xmath37 ( @xmath38 ) has only one real root @xmath39 ( @xmath40 ) given by @xmath41^{1/3 } + \left [ \frac{1}{2 } \left ( \pm h_0 - \sqrt{h_0 ^ 2 - h_c^2}\right ) \right]^{1/3 } , \label{eq : define - s_pm}\ ] ] which are indicated in fig .
[ fig : f_x](b ) . figures ( a ) and ( b ) show the motions obtained by numerically integrating eq .
( [ eq : tdgl - external - field ] ) for @xmath42 and @xmath43 , where two sbm s ( solid line ) and an unstable srm ( dashed line ) are drawn .
the unstable srm is evaluated by replacing @xmath44 .
on the other hand , figs .
( c ) and ( d ) show the motions for @xmath45 and @xmath46 . ]
next let us consider the dynamics described by eq .
( [ eq : tdgl - external - field ] ) for @xmath28 and for @xmath30 , and discuss similarity and difference between the dynamics in the periodically oscillating field case and those in the present dmn case .
a part of our results belongs to the context of the noise - induced phase transition and mfpt in refs . @xcite . in the case of @xmath28 ,
three motions numerically integrated are shown in figs .
[ fig : orb](a ) and ( b ) .
two motions confined in the ranges @xmath47 and @xmath48 are both stable .
the long time average @xmath49 of each motion does not vanish , and the motion is called sbm in relation to dpt in the oscillating external field case . on the other hand ,
the motion @xmath50 confined in the range @xmath51 is unstable .
the long time average of @xmath50 vanishes , and in this sense the motion is called srm .
it should be noted that this unstable srm is located between two stable sbm , which has a similar characteristic to sbo of dpt @xcite .
the motion of @xmath8 for @xmath30 is shown in figs .
[ fig : orb](c ) and ( d ) .
one observes that there exists a stable srm confined in the range @xmath52 . for srm ,
the time average of @xmath8 vanishes , i.e. , @xmath53 .
the comparison between figs .
[ fig : orb](b ) and ( d ) suggests that the srm for @xmath54 is generated via the `` attractor merging crisis '' @xcite of the two sbm s and one unstable srm , i.e. , the two sbm s and one unstable srm disappear and then one stable srm takes place at @xmath55 .
this situation is similar to that in the dpt case .
however , in contrast to the dpt case , as will be shown in sec .
[ sec : stationary - distribution ] , the transition line on the ( @xmath56 , @xmath57 ) plane is independent of the correlation time @xmath58 of @xmath9 and the average @xmath49 depends discontinuously on @xmath57 . in this subsection , we discuss the stationary distribution functions for sbm and srm . to this aim
, we first consider a slightly general nonlinear langevin equation of motion driven by dmn , @xmath59 where @xmath60 and @xmath61 are generally nonlinear functions of @xmath62 and @xmath9 is dmn @xcite .
the temporal evolution of the distribution function @xmath63 that @xmath64 and @xmath9 respectively take the values @xmath62 and @xmath65 is determined by @xcite @xmath66 , \nonumber \\
\frac{\partial}{\partial t}q(x , t ) & = & - \frac{2}{\tau_f } q(x , t ) - \frac{\partial}{\partial x } \left [ f(x)q(x , t ) + h_0 g(x)p(x , t ) \right ] , \label{eq : general - probability - evolution}\end{aligned}\ ] ] where we put @xmath67 and @xmath68 .
the stationary distribution @xmath69 is solved to yield @xmath70 \right\ } , \label{eq : general - stationary - distribution}\ ] ] provided that each of the equations @xmath71 has at least one stable fixed point , where @xmath72 is the normalization constant . by substituting @xmath73 and @xmath74 , ( eq .
( [ eq : tdgl - external - field ] ) ) , into eq . ( [ eq : general - stationary - distribution ] ) , the stationary distribution function @xmath75 for sbm ( @xmath28 ) for @xmath76 or @xmath77 is written as @xmath78 where @xmath79 , ( 2,3,1 ) , and ( 3,1,2 ) .
stationary distribution functions obtained both analytically ( solid line ) and numerically ( symbols ) for ( a ) @xmath80 and @xmath43 , and for ( b ) @xmath81 , 0.4 , and 0.5 by keeping @xmath46 .
( a ) and ( b ) correspond respectively to sbm and srm .
the distribution only in the range @xmath47 is drawn in ( a ) .
the analytical result is in good agreement with that of the numerical simulation . ] on the other hand , the stationary distribution function @xmath82 for the srm ( @xmath30 ) for @xmath83 is obtained as @xmath84^ { - \frac{\tau_f^{-1}}{3 s_{+}^2 -1 } - 1 } \nonumber \\ & \times & \exp \bigg\{\frac{\tau_f^{-1 } s_{+}}{(s_{+}^2 - 1/3)\sqrt{3s_{+}^2 - 4 } } \bigg [ \arctan \left ( \frac{2 s - s_{+}}{\sqrt{3 s_{+}^2 - 4 } } \right ) \nonumber \\ & & - \arctan \left ( \frac{2 s + s_{+}}{\sqrt{3 s_{+}^2 - 4 } } \right ) \bigg ] \bigg\}. \label{eq : distribution - srm}\end{aligned}\ ] ] the analytic solutions ( [ eq : distribution - sbm ] ) and ( [ eq : distribution - srm ] ) are numerically confirmed in fig .
[ fig : distrib ] . as @xmath57
is increased , the form of the stationary distribution function changes drastically from the forms in eq .
( [ eq : distribution - sbm ] ) to eq .
( [ eq : distribution - srm ] ) at @xmath55 .
this phenomenon which is induced by the disappearance of two pairs of stable and unstable fixed points @xcite is an example of the noise - induced phase transitions @xcite .
phase diagram on the ( @xmath56 , @xmath57 ) plane .
transition line is independent of @xmath56 and is given by @xmath55 . ]
it turns out that the transition line between srm and sbm on the ( @xmath56 , @xmath57 ) plane is given by @xmath55 . the phase diagram is given in fig .
[ fig : phase_diagram ] .
furthermore , the long time average of @xmath8 , @xmath49 , depends discontinuously on @xmath57 at @xmath55 as shown in fig .
[ fig : disconnect ] .
these behaviors are quite different from those of the dpt case driven by periodically oscillating field , @xmath85 @xcite .
the transition point @xmath86 for a fixed @xmath0 between sro and sbo depends on the frequency @xmath0 , and @xmath49 is a continuous function of @xmath1 .
@xmath87 vs @xmath57 for @xmath43 .
one finds that @xmath87 is discontinuous at @xmath55 . ]
we hereafter discuss the dynamics for @xmath57 slightly above @xmath32 .
let us first consider the behavior obeying the equations @xmath88 for @xmath30 , i.e. , @xmath9 is fixed to be either @xmath18 or @xmath19 .
equation ( [ eq : tdgl - fixed - field ] ) for @xmath30 is integrated to yield @xmath89 , \label{eq : tdgl - srm - motion}\end{aligned}\ ] ] where @xmath90 and @xmath91 has been defined in eq .
( [ eq : define - s_pm ] ) .
three orbits of the equation of motion ( [ eq : tdgl - fixed - field ] ) with @xmath92 for @xmath30 .
the values of @xmath57 are set to be @xmath93 ( dotted line ) , 0.4 ( dashed line ) , and 0.5 ( solid line ) , where all the initial conditions are chosen as @xmath94 . ]
figure [ fig : anal_sro_orbit ] displays three orbits given by eq . ( [ eq : tdgl - srm - motion ] ) with @xmath94 and @xmath92 , which shows that @xmath8 approaches @xmath39 in the limit @xmath95 .
one observes that @xmath8 stays for a long time in the vicinity of @xmath96 for @xmath57 slightly above @xmath32 .
the small region including the position @xmath97 is called the ` channel ' . from the symmetry of the system , there also exists the channel near @xmath98 for @xmath99 , as shown in fig .
[ fig : f_x](b ) .
let us express the positions @xmath100 of the channels as @xmath101 the characteristic time @xmath102 is then defined as the time span that the state point @xmath8 passes through one of the channels for a constant @xmath9 , either @xmath103 or @xmath104 .
@xmath102 can be estimated by integrating eq .
( [ eq : tdgl - fixed - field ] ) around @xmath105 as follows .
first , consider the case @xmath106 . by setting @xmath107 and assuming @xmath108 , eq .
( [ eq : tdgl - fixed - field ] ) is approximated as @xmath109 this can be integrated to give @xmath110 \label{eq : channel - motion}\ ] ] with the initial condition @xmath111 .
@xmath102 is estimated by the condition @xmath112 and thus @xmath113 let us next consider the process that the state point @xmath8 passes through the channels under dmn .
figure [ fig : time - series ] shows temporal evolutions of @xmath8 numerically obtained for @xmath114 and 0.385 .
one finds that the time of passing through channels increases as @xmath57 approaches @xmath32 .
the mfpt @xmath115 through channels was calculated in refs .
@xcite by analyzing the master equation . in the present section
, we will derive mfpt in terms of the time scales @xmath58 and @xmath102 from a phenomenological viewpoint without use of the analysis made in refs .
@xcite .
time series of @xmath8 for ( a ) @xmath116 and ( b ) 0.385 by keeping @xmath117 . ]
the condition for passing through a channel is that @xmath9 continues to take the identical value either @xmath18 or @xmath104 for time longer than @xmath102 . for @xmath57 satisfying @xmath118 , we obtain @xmath119 , which implies that @xmath9 almost always satisfies the condition for passing through the channel . therefore , @xmath115 in the case of @xmath118 is nearly equivalent to @xmath102 , i.e. , @xmath120 in the case of @xmath121 , on the other hand , eq . ( [ eq : field - probability ] ) gives @xmath122 .
this fact implies that the probability that @xmath9 continues to take the identical value for time longer than @xmath102 is quite small and hence that @xmath115 is much longer than @xmath102 because it needs a long time to satisfy the condition for the state point to pass through the channel .
@xmath115 in the case of @xmath121 is explicitly determined as follows . for a long @xmath115 ,
let us divide @xmath115 into subintervals each of which has the time span @xmath58 .
the divided individual time series are approximately independent of each other . therefore , @xmath123 is the probability that the state point passes through a channel once because @xmath115 is mfpt through the channel . on the other hand
, @xmath124 is identical to the probability for @xmath8 to pass through the channel once by definition of the probability .
therefore we get the relation @xmath125 , which leads to @xmath126 \label{eq : estimate - tau - vs - h0}\ ] ] with the constant @xmath127 defined in eq .
( [ eq : tau_channel ] ) .
this expression agrees with the result obtained in refs .
@xcite .
the dependence of @xmath115 on @xmath128 determined numerically for @xmath43 .
@xmath115 for a given @xmath57 is obtained by integrating eq .
( [ eq : tdgl - external - field ] ) until a time @xmath3 and simultaneously by counting the number @xmath72 of times passing through the position @xmath129 .
then @xmath115 is evaluated as @xmath130 .
one finds that @xmath131 linearly depends on @xmath132 in the region of @xmath133 , where the solid line is expressed as @xmath134 with fitting coefficients @xmath135 and @xmath136 . ]
equation ( [ eq : estimate - tau - vs - h0 ] ) reveals that mfpt through the channel depends on @xmath137 in a stretched exponential form for @xmath121 , and is quite different from the asymptotic form ( [ eq : average - time - far - point ] ) .
the above dependence of @xmath115 on @xmath128 is confirmed in fig .
[ fig : tau_h0 ] .
time series of @xmath9 schematically indicating the time that @xmath8 jumps from @xmath138 to @xmath139 .
@xmath8 jumps at the time @xmath140 in this case , where @xmath9 takes the value @xmath19 longer than @xmath102 for the first time in the time series of @xmath9 . ] in order to discuss statistical characteristics of the dynamics passing through the channels for @xmath121 , we here develop a phenomenological approach .
the behaviors of @xmath8 for which we attempt to model are first summarized .
the initial condition of @xmath8 is set to be in the vicinity of @xmath138 .
if a time interval of @xmath9 satisfying the condition @xmath141 becomes longer than @xmath102 for the first time , then @xmath8 passes through @xmath100 and approaches @xmath139 in the time interval .
see fig .
[ fig : jump_schematic ] .
the event in which @xmath8 jumps from @xmath138 to @xmath139 occurs only in this case .
it should be noted that the jumps from @xmath142 ( @xmath143 ) to @xmath143 ( @xmath142 ) are approximately independent of subsequent jumps .
let us discretize the time @xmath6 in the form @xmath144 , ( @xmath145 , 2 , 3 , @xmath146 ) as a simple approach to develop the phenomenological analysis according to the process noted above , where @xmath147 is a certain small time step .
then , @xmath102 is discretized as @xmath148 with the corresponding integer @xmath149 .
@xmath9 is assumed to keep the same value for the interval @xmath147 , which is denoted as @xmath150 .
the conditional probability @xmath151 that @xmath152 takes the same value as @xmath153 is given by @xmath154 and the probability @xmath155 that @xmath156 is different from @xmath157 is therefore given by @xmath158 the system is analyzed phenomenologically as follows : * we introduce the variable @xmath159 at a discretized time @xmath160 which takes two values @xmath24 .
* @xmath159 and @xmath157 are initially set to @xmath161 and @xmath162 , respectively . * @xmath159 jumps from @xmath163 ( @xmath164 ) to @xmath164 ( @xmath163 ) only if @xmath157 continues to take the identical value @xmath19 ( @xmath18 ) for a time interval longer than @xmath165 .
* @xmath159 does not jump from @xmath163 ( @xmath164 ) to @xmath164 ( @xmath163 ) even though @xmath157 continues to take @xmath103 ( @xmath104 ) for any time interval longer than @xmath165 .
we first derive the exact expression for the mfpt @xmath115 through the channel with the phenomenological approach . in considering the time series having @xmath157 ,
@xmath115 is evaluated as @xmath166 here @xmath167 is the number of the time sequences @xmath168 for @xmath169 satisfying that @xmath153 changed its value @xmath170 times in each @xmath168 and @xmath171 jumps from @xmath163 to @xmath164 for the first time at @xmath172 . equation ( [ eq : average - tau ] ) is , furthermore , rewritten as @xmath173 with the differential operator @xmath174 and the quantity @xmath175 defined by @xmath176 and @xmath177 one should note that the @xmath155- and @xmath151-dependences in @xmath178 are crucial and that @xmath155 and @xmath151 are considered to be independent in eq .
( [ eq : q - q - p ] ) . the explicit form of @xmath175
is then determined so as to satisfy the following conditions : * in considering any length of time series giving @xmath157 , there exists a time interval of length @xmath149 in the last of the time series , where all the @xmath157 take the same value @xmath19 , i.e. , the condition that @xmath159 jumps from @xmath163 to @xmath164 is satisfied . *
the condition for @xmath159 to jump from @xmath163 to @xmath164 is not satisfied before the last time interval .
one should note that the equality @xmath179 holds for any @xmath180 , because the time interval described above always exists somewhere in a long time series .
particularly , for @xmath181 , @xmath182 is obviously equal to the probability that @xmath171 changes its sign , which must be unity for @xmath30 .
as shown in appendix [ sec : form - qn ] , the explicit form of @xmath183 is given by @xmath184 where the condition @xmath179 is easily confined . applying the operator ( [ eq : t - operator ] ) to the explicit form ( [ eq : explicit - form - qn ] ) with @xmath181 yields the relation @xmath185 where the last equality is obtained by using eqs .
( [ eq : explicit - p ] ) and ( [ eq : explicit - q ] ) with the relation @xmath186 . the exact expression of @xmath115 is finally given by @xmath187 in the limit of @xmath188 by keeping @xmath102 constant . equation ( [ eq : solution - tau - vs - h0 ] ) qualitatively agrees for @xmath189 with the result ( [ eq : estimate - tau - vs - h0 ] ) .
the distribution function @xmath190 for the passage time @xmath20 through the channel @xmath100 is determined by solving the equation @xmath191 where @xmath192 is the delta function .
the laplace transform @xmath193(z)$ ] should be calculated in order to solve eq .
( [ eq : distribution - func - tau ] ) . by using the series expansion of @xmath175 given by eq .
( [ eq : q - q - p ] ) , the laplace transform of @xmath190 is obtained as @xmath194(z ) & \equiv & \int_0^{\infty } e^{-\tau z } p(\tau ) d\tau = \left .
e^{-z \hat{t } } q_{n_{ch}}(q , p ) \right|_{q=1-p } \nonumber \\ & = & \left .
\sum_{l } \sum_{0 \le k \le l } g_{k , l}^{(n_{ch } ) } ( e^{-z \delta t } q)^k ( e^{-z \delta t } p)^{l - k } \right|_{q=1-p}. \label{eq : formal - laplace - p}\end{aligned}\ ] ] equation ( [ eq : formal - laplace - p ] ) implies that @xmath193(z)$ ] can be obtained by replacing @xmath155 and @xmath151 by @xmath195 and @xmath196 in @xmath175 , respectively , i.e. , @xmath197(z ) = \left .
q_{n_{ch } } ( e^{-z \delta t } q , e^{-z \delta t } p ) \right|_{q=1-p}. \label{eq : explicit - laplace - p}\ ] ] substituting the explicit form ( [ eq : explicit - form - qn ] ) for @xmath181 into eq .
( [ eq : explicit - laplace - p ] ) yields the equation @xmath197(z ) = \frac { ( \tau_f z + 1)e^{-(z + \tau_f^{-1})\tau_{ch } } } { \tau_f^2 z^2 + 2 \tau_f z + e^{-(z + \tau_f^{-1 } ) \tau_{ch } } } \label{eq : exact - laplace - p}\ ] ] in the limit of @xmath188 by keeping @xmath102 constant . by applying the inverse laplace transform to eq .
( [ eq : exact - laplace - p ] ) , the distribution function @xmath190 is analytically evaluated in the series expansion as @xmath198 , \label{eq : distribution - p}\end{aligned}\ ] ] where @xmath199 and @xmath200 is the heaviside function defined by @xmath201 for details of the derivation of eq .
( [ eq : distribution - p ] ) , see appendix [ sec : form - pt ] .
let us suppose to truncate the expansion ( [ eq : distribution - p ] ) at @xmath202 for an arbitrary @xmath203 .
it should be noted that eq .
( [ eq : distribution - p ] ) gives the exact distribution for @xmath204 even though the truncation is executed , since all the terms individually include @xmath205 and so the terms for @xmath206 do not contribute to @xmath190 for @xmath207 .
the analytical result ( [ eq : distribution - p ] ) is compared with the numerically evaluated distribution in fig .
[ fig : distrib_ptau ] .
one observes that the phenomenological analysis quantitatively explains the statistical property of passing through the channels .
the characteristics obtained from the figure are summarized as follows .
* there exists a region where @xmath208 for @xmath209 , which presents the minimal time of passing through the channels . *
@xmath190 decreases exponentially for @xmath210 , @xmath211 with a constant @xmath212 . *
the rate @xmath212 increases as @xmath58 is increased .
this tendency is consistent with the fact that the probability of passing through channels increases as @xmath58 is increased since dmn will often continue to take an identical value longer than @xmath102 .
distribution function @xmath190 of time passing through channels .
the theoretical results ( solid lines ) are compared with the numerically evaluated distributions ( symbols ) for ( a ) @xmath213 and ( b ) @xmath214 .
the strength of dmn is set to be @xmath215 in both of the numerical simulations .
@xmath102 corresponding to the applied dmn is estimated to be @xmath216 via the condition @xmath217 for @xmath209 , where both ( a ) and ( b ) give the identical value of @xmath102 . the expansion in eq .
( [ eq : distribution - p ] ) is summed over for @xmath218 , i.e. , @xmath219 . ] on the other hand , the expansion ( [ eq : distribution - p ] ) disagrees with the correct value in an exponential way for @xmath220 . let us try to obtain the asymptotic solution of @xmath190 for @xmath210 . equation ( [ eq : laplace - expand ] ) is approximated as @xmath221 where @xmath115 is mfpt given in eq .
( [ eq : solution - tau - vs - h0 ] ) .
the inverse laplace transform of eq .
( [ eq : approximate - laplace - p ] ) is straightforwardly calculated to give @xmath222 which reveals that @xmath190 decreases exponentially with the damping rate @xmath223 for @xmath210 .
we derive the fourier spectrum of a time series @xmath8 by the phenomenological analysis to focus on the dynamical characteristics in the srm phase .
the fourier spectrum @xmath224 is defined by @xmath225 i.e. , the ensemble average of the fourier transform of a time series @xmath64 .
let us first consider @xmath226 $ ] .
then the time series @xmath227 is expressed as @xmath228 with @xmath229 , where @xmath230 denotes the @xmath180th time to cross zero for @xmath8 .
hereafter , @xmath231 is set to be zero without loss of generality . by identifying that @xmath232
is independently distributed according to eq .
( [ eq : distribution - p ] ) , one obtains the fourier spectrum of @xmath227 by the phenomenological analysis shown in appendix [ sec : structure - func ] , in the form @xmath233 where @xmath234 represents the real part of @xmath235 , and @xmath236 is used . substituting the explicit form of @xmath237 given in eq .
( [ eq : exact - laplace - p ] ) with @xmath238 into eq .
( [ eq : general - structure - function ] ) yields @xmath239 the above result is confirmed by comparing with the numerically evaluated fourier spectrum for the normalized time series @xmath227 in fig .
[ fig : power_normal ] .
fourier spectrum of @xmath227 obtained theoretically ( solid line ) and numerically ( dashed line ) .
@xmath240 is set for both numerical and theoretical results , and the strength of dmn is set to be @xmath241 in the numerical simulation .
@xmath242 estimated in fig .
[ fig : distrib_ptau ] is used for eq .
( [ eq : get - structure - function ] ) . ]
let us finally modify the phenomenological analysis which is compatible with the numerically evaluated spectrum of the original time series @xmath8 without normalization . instead of eq .
( [ eq : orbit - probability - model ] ) , let us define @xmath243 \quad \mbox{for \(t_{n-1 } \le t < t_{n}\)}\ ] ] with @xmath229 , where @xmath244 incorporates the wave form of the time series passing through the channel and is assumed to be @xmath245 for @xmath246 .
note that by setting @xmath245 also for @xmath247 the result of original phenomenological analysis is recovered .
as shown in appendix [ sec : structure - func ] , the fourier spectrum @xmath248 for @xmath249 as a modification to @xmath250 is obtained in the form @xmath251}{4 } , \label{eq : correct - structure - function}\ ] ] where @xmath252 by approximating as @xmath253 for @xmath254 , eq .
( [ eq : correct - structure - function ] ) reduces to @xmath255 theoretically obtained fourier spectrum ( [ eq : explicit - correct - structure - function ] ) ( solid line ) with @xmath256 is compared with the numerical result ( dashed line ) .
@xmath240 is set for both numerical and theoretical results , and the strength of dmn is set to be @xmath241 in the numerical simulation .
@xmath242 is used for eq .
( [ eq : get - structure - function ] ) which was estimated in fig .
[ fig : distrib_ptau ] . ] equation ( [ eq : explicit - correct - structure - function ] ) is confirmed by comparing with the numerically evaluated fourier spectrum in fig .
[ fig : power_full ] .
in this paper , we used the model equation ( [ eq : tdgl - external - field ] ) under the dichotomous markov noise ( dmn ) @xmath9 with a finite correlation time in order to investigate the dynamics of the magnetization @xmath8 of the ferromagnet system driven by the magnetic field applied in one direction , where its strength is constant and only the direction temporally changes .
it was found that the dynamics of @xmath8 show two kinds of motion , i.e. , the symmetry - restoring motion ( srm ) and the symmetry - breaking motion ( sbm ) , which are respectively observed when @xmath57 is above and below the critical value @xmath32 .
the transition line between srm and sbm was determined only by the strength of the applied dmn and is independent of the correlation time @xmath58 of dmn . by observing the distribution functions of @xmath8 for srm and sbm ,
the ensemble average of @xmath8 discontinuously changes at @xmath55 .
these results are quite different from those in the system driven by a periodically oscillating field @xcite .
we then discussed the mean first passage time ( mfpt ) slightly above @xmath32 and found that it depends on @xmath137 and @xmath58 as @xmath257.\ ] ] this anomalous characteristics has a form similar to that of the average duration between neighboring phase slips of the phase difference in the phase synchronization observed in coupled chaotic systems @xcite .
furthermore , a phenomenological approach was proposed to analytically discuss the statistical characteristics for @xmath30 . by obtaining the probability @xmath183 of @xmath8 passing through the channel ,
the mfpt @xmath115 , the distribution function @xmath190 of the passage time @xmath20 and the fourier spectrum @xmath258 of the time series were obtained by using the phenomenological analysis .
the statistics obtained by the phenomenological analysis were found to be not only in qualitative but also in quantitative agreement with the numerically obtained results . in closing the paper ,
it is worth noting that the effect of dmn on nonlinear dynamical systems is generally quite different from that of gaussian noise and dmn produces a new dynamical response of the systems .
it is highly desired to examine the statistical characteristics obtained in this paper in laboratory experiments , and also in other numerical simulations , e.g. , monte carlo simulations .
the authors thank n. tsukamoto , n. fujiwara , and h. hata for valuable comments .
this study was partially supported by grant - in - aid for scientific research ( c ) of the ministry of education , culture , sports , science , and technology , and the 21st century coe program `` center of excellence for research and education on complex functional mechanical systems '' at kyoto university .
first of all , notice that each sample path of @xmath157 ( @xmath259 ) corresponds to a symbol sequence of @xmath260 .
let @xmath261 be a symbol sequence of @xmath262 , which is referred to as a string .
for given sets @xmath135 and @xmath136 of strings , let @xmath263 be the set of all strings expressed as @xmath264 for @xmath265 and @xmath266 , where @xmath264 denotes the concatenated string of @xmath267 followed by @xmath268 . in the following ,
the set composed of only one string @xmath261 will be simply expressed as @xmath261 .
furthermore , let @xmath269 be the set of all strings expressed as @xmath270 with @xmath271 and @xmath272 , where @xmath273 means the zero length string , including the case @xmath274 for @xmath275 . then , with this notation , every possible sequence starting with @xmath276 and terminating with successive @xmath277 s of length @xmath180 @xmath278 , which firstly appears in that string , can be expressed as @xmath279 for a given set @xmath280 of strings , let us define its `` probability '' as a function of @xmath151 and @xmath155 by @xmath281 where @xmath282 denotes the probability for a string @xmath261 , and @xmath283 and @xmath284 denote the numbers of pairs of identical symbols ( @xmath285 , @xmath286 ) and different symbols ( @xmath287 , @xmath288 ) appearing in @xmath261 , respectively . for given strings @xmath289 and @xmath290 , obviously , the identity @xmath291 holds with @xmath292 being the first symbol in @xmath290 , and we call @xmath293 and @xmath290 a decomposition of the string @xmath294 . by noting that @xmath295 and considering decompositions of each element in @xmath296 , @xmath297 can be expressed as @xmath298 where @xmath299 and @xmath300 since each element of @xmath301 can be uniquely decomposed into a multiple of elements in @xmath302 and inversely every multiple of elements in @xmath303 uniquely corresponds to an element in @xmath301 as its decomposition , we obtain @xmath304^j\\ \label{eq : p - s_n } & = & \frac{1}{1-p(q , p ; s''_n)}.\end{aligned}\ ] ]
finally , @xmath305 and @xmath306 are calculated as @xmath307 and @xmath308 which yield eq .
( [ eq : explicit - form - qn ] ) together with ( [ eq : q_n ] ) and ( [ eq : p - s_n ] ) .
let us evaluate the inverse laplace transform of eq .
( [ eq : exact - laplace - p ] ) in order to derive the explicit form of @xmath190 . for simplicity , the time is rescaled as @xmath309 .
then eq .
( [ eq : exact - laplace - p ] ) is rewritten as @xmath310 which reads @xmath311= \langle e^{-z(\tau - \tau_{ch } ) } \rangle = \frac{e^{-\tau_{ch}}}{1+z - \frac{1-e^{-(1+z)\tau_{ch}}}{1+z } } = e^{-\tau_{ch } } \sum_{n=0}^{\infty } \frac{\left(1-e^{-(1+z)\tau_{ch } } \right)^n } { ( 1+z)^{2 n+1}}. \label{eq : laplace - expand}\ ] ] repeatedly applying the formula @xmath312
= e^{-\tau_{ch } } f(\tau-\tau_{ch } ) = e^{-\tau_{ch } } e^{- \tau_{ch}\frac{d}{d\tau } } f(\tau ) = e^{-\tau_{ch } } e^{- \tau_{ch}\frac{d}{d\tau } } { \cal l}^{-1}[\hat{f}(z)]\ ] ] with @xmath313 $ ] , one obtains @xmath314 & = & ( 1-e^{-\tau_{ch}}e^{-\tau_{ch}\frac{d}{d\tau}})^n { \cal l}^{-1}\left[\frac{1}{(1+z)^{2n+1}}\right ] \nonumber \\ \nonumber \\ & = & ( 1-e^{-\tau_{ch}}e^{-\tau_{ch}\frac{d}{d\tau}})^n \frac{\tau^{2n}}{(2n)!}e^{-\tau}\theta(\tau),\end{aligned}\ ] ] where @xmath315 denotes the heaviside function eq .
( [ eq : heaviside ] ) , and the formula @xmath316 = \frac{\tau^{m-1}}{(m-1 ) ! } e^{-\tau } \theta(\tau)\ ] ] for any positive integer @xmath317 was used .
thus , the inverse laplace transform of eq .
( [ eq : laplace - expand ] ) reads @xmath318 where the identity @xmath319 was applied . by noting the identity @xmath320 eq .
( [ eq : p - tau-1 ] ) is further simplified as @xmath321 where @xmath322 and the formula @xmath323 was used .
after the replacement @xmath324 and the rescaling of time as @xmath325 and @xmath326 , eq .
( [ eq : p - tau - fin ] ) reduces to eq .
( [ eq : distribution - p ] ) .
in this appendix , we derive the fourier spectrum of the time series of the magnetization @xmath8 for the phenomenological analysis .
let @xmath327 be a sequence of mutually independent random variables having the probability density @xmath190 obeying the condition @xmath208 for @xmath328 , i.e. , @xmath329 .
we introduce the variable @xmath330 \mbox { for } t_{n-1}\le t < t_n,\ ] ] where @xmath331 and @xmath244 is a function satisfying @xmath245 for @xmath332 .
we assume that the time series of the magnetization @xmath8 is approximately expressed by @xmath249 with an appropriate form of @xmath333 .
the fourier transform of @xmath249 follows @xmath334e^{-i\omega t}dt \nonumber \\ & = & \sum_{n=1}^n(-1)^{n-1}e^{-i\omega t_{n-1}}\left [ \int_0^{\tau_n}e^{-i\omega t}dt-\int_0^{\tau_{ch}}a(t)e^{-i\omega t}dt \right ] \nonumber \\ & = & \sum_{n=1}^n(-1)^{n}e^{-i\omega t_{n-1}}\frac{e^{-i\omega \tau_n } -\hat{a}(\omega)}{i\omega } , \label{eq : fourier - probability - model}\end{aligned}\ ] ] where @xmath335 considering the absolute square of eq .
( [ eq : fourier - probability - model ] ) and then taking the ensemble average , we obtain @xmath336 \label{eq : absolute - fourier - probability}\end{aligned}\ ] ] and @xmath337 } \nonumber\\ & & + 2n^{-1}\re\left[\sum_{1 \le m < n \le n}(-1)^{n - m } \langle e^{-i\omega\tau_k}\rangle^{n - m-1 } \left(\langle e^{-i\omega\tau_k}\rangle-\hat{a}(\omega)\right ) \left(1-\langle e^{-i\omega\tau_k}\rangle \hat{a}^*(\omega)\right ) \right ] .
\label{eq : fourier-1}\end{aligned}\ ] ] since @xmath338 , in the limit of @xmath339 , the last term in eq .
( [ eq : fourier-1 ] ) reads @xmath340 } \nonumber\\ & & = -2\re\left[\frac{\langle e^{-i\omega\tau_n}\rangle \left(1+|\hat{a}(\omega)|^2\right)- \hat{a}(\omega)-\langle e^{-i\omega\tau_n}\rangle^{2 } \hat{a}^*(\omega)}{1+\langle e^{-i\omega\tau_n}\rangle}\right],\end{aligned}\ ] ] which leads to @xmath341\left ( 1+|\hat{a}(\omega)|^2 + 2\re[\hat{a}(\omega)]\right).\ ] ] noting that @xmath236 , we obtain @xmath342}{4},\ ] ] where @xmath343 \label{eq : power - spectrum-1}\ ] ] corresponds to the case that @xmath245 for all @xmath344 and thus @xmath345 for @xmath346 .
99 t. tom and m. j. de oliveira , phys .
rev . a * 41 * , 4251 ( 1990 ) ; j. f. f. menders , and e. j. s. large , j. stat . phys . * 64 * , 653 ( 1991 ) .
w. s. lo and r. a. pelcovits , phys .
a * 42 * , 7471 ( 1990 ) m. acharyya , phys .
e , * 56 * 1234 ( 1997 ) ; * 56 * , 2407 ( 1997 ) ; * 58 * , 179 ( 1998 ) ; * 59 * , 218 ( 1999 ) ; s. w. sides , p. a. rikvold , and m. a. novotny , phys . rev .
lett . * 81 * , 834 ( 1998 ) ; phys .
e * 59 * , 2710 ( 1999 ) .
p. a. rikvold _
et al_. , in _
computer simulation studies in condensed matter physics xiii _ , edited by d. p. landau , s. p. lewis , and h .- b .
schttler ( springer , berlin , 2000 ) , pp . 105 - 119 .
q. jiang , h .-
yang , and g .- c .
wang , phys .
b * 52 * , 14911 ( 1995 ) ; j. appl .
* 79 * , 5122 ( 1996 ) .
n. g. van kampen , _
stochastic processes in physics and chemistry _
( north - holland , amsterdam , 1981 ) .
k. kitahara , w. horsthemke , and r. lefever , phys .
* a70 * , 377 ( 1979 ) ; k. kitahara , w. horsthemke , r. lefever , and y. inaba , prog . theor .
phys . * 64 * , 1233 ( 1980 ) . i. bena , c. van den broeck , r. kawai , and k. lindenberg , phys .
e * 66 * , 045603(r ) ( 2002 ) ; i. bena , c. van den broeck , r. kawai , and k. lindenberg , _ ibid _ * 68 * , 041111 ( 2003 ) . i. lheureux and r. kapral , j. chem .
phys . * 88 * , 7468 ( 1988 ) ; * 90 * , 2453 ( 1989 ) ; j. m. porr , j. masoliver , and k. lindenberg , phys .
a * 44 * , 4866 ( 1991 ) ; j. m. porr , j. masoliver , k. lindenberg , i. lheureux , and r. kapral , _ ibid _ * 45 * , 6092 ( 1992 ) .
p. hnggi and p. riseborough , phys .
a * 27 * , 3379 ( 1983 ) ; c. van den broeck and p. hnggi , _ ibid _ * 30 * , 2730 ( 1984 ) ; j. m. sancho , _ ibid _ * 31 * , 3523 ( 1985 ) ; j. masoliver , k. lindenberg , and b. j. west , _ ibid _ * 34 * , 2351 ( 1986 ) ; m. a. rodriguez and l. pesquera , _ ibid _ * 34 * , 4532 ( 1986 ) .
w. horsthemke and r. lefever , in _ noise - induced transitions _
( springer , berlin , 1984 ) , in chap . 9
. i. bena , arxiv : cond - mat/0606116 e. ott , _ chaos in dynamical systems _
( university press , cambridge , 1993 ) , sec . 8.3 . there
exist several techniques in the case that @xmath9 is a white noise .
see e.g. , c. w. gardiner , _ handbook of stochastic methods _ ( springer , berlin , 1983 ) ; h. risken , _ the fokker - planck equation _ ( springer , berlin , 1984 ) .
see e.g. , m. g. rosenblum , a. s. pikovsky and j. kurths , phys .
* 76 * , 1804 ( 1996 ) ; a. pikovsky , g. osipov , m. rosenblum , m. zaks , and j. kurths , phys .
lett . * 79 * , 47 ( 1997 ) ; s. boccaletti , e. allaria , r. meucci , and f. t. arecchi , phys .
lett . * 89 * , 194101 ( 2002 ) ; e. rosa , e. ott , and m. h. hess , phys .
80 * , 1642 ( 1998 ) ; k. j. lee , y. kwak , and t. k. lim , phys . rev .
lett . * 81 * , 321 ( 1998 ) ; h. fujisaka , t. yamada , g. kinoshita , and t. kono , physica d * 205 * , 41 ( 2005 ) ; h. fujisaka , s. uchiyama , and t. horita , prog .
phys . * 114 * , 289 ( 2005 ) . | an ising spin system under the critical temperature driven by a dichotomous markov noise ( magnetic field ) with a finite correlation time is studied both numerically and theoretically .
the order parameter exhibits a transition between two kinds of qualitatively different dynamics , _ symmetry - restoring _ and _ symmetry - breaking motions _ , as the noise intensity is changed
. there exist regions called channels where the order parameter stays for a long time slightly above its critical noise intensity . developing a phenomenological analysis of the dynamics , we investigate the distribution of the passage time through the channels and the power spectrum of the order parameter evolution .
the results based on the phenomenological analysis turn out to be in quite good agreement with those of the numerical simulation . |
cataclysmic variables are spectrally exotic binaries , displaying a diversity and complexity of features , behavior , and variability , effectively masking the spectral signatures of their coolest secondaries . despite unyielding efforts to unambiguously and directly detect likely
substellar companions in either non - magnetic or magnetic cataclysmic variables , these altered degenerate dwarfs remain largely elusive @xcite .
the mass determination of the substellar donors in two non - magnetic accreting binaries reported by @xcite and @xcite remain rare among cataclysmic variables , although the determination of their substellarity was indirect .
cataclysmic variables containing strongly magnetic white dwarfs ( called polars because of the highly polarized nature of their light ) offer a slight advantage by lacking the often overwhelming luminosity from an accretion disk .
on the other hand , their magnetic fields generate copious cyclotron radiation ( in addition to flux from direct accretion ) , which is not present in non - magnetic cataclysmic variables .
polars are also known for exhibiting long periods of quiescence in which many secondaries have been directly observed , although none of spectral type later than m @xcite
. the recently recognized class of low accretion rate polars may provide decent infrared hunting grounds for substellar companions , if present , as their cyclotron harmonics appear mainly in the optical due to @xmath3 mg fields .
it has been suggested that no mass transfer takes place in these systems , but rather the efficient capture of secondary wind , and that their separations are consistent with being essentially detached @xcite .
yet , it is difficult to certify that these binaries are not simply in protracted low states such as seen in other polars @xcite .
recent _ spitzer _ irac observations of several polars indicate flat infrared spectral energy distributions which are difficult to reconcile with cool degenerate companions alone , and may indicate circumbinary dust as well as underlying cyclotron continuum emission @xcite .
it will be interesting to see the results of similar mid - infrared observations with a larger sample of polars as well as observations at longer wavelengths where there should be little or no cyclotron emission even in @xmath4 mg fields .
the first published _
irs spectrum of a polar strongly supports the presence of circumbinary dust at ef eridani @xcite .
this paper represents a spectroscopic effort to directly detect the almost certain substellar companion to the cool magnetic white dwarf sdss j121209.31 + 013627.7 ( sdss 1212 ) .
its binarity was reported with optical spectra supplemented with a single @xmath5-band photometric observation , limiting the spectral type of the secondary to l5 or later , and was initially thought to represent a detached system @xcite .
this possibility was still considered viable despite a second study which detected strong cyclotron emission a likely indicator of mass transfer in @xmath6-band time series photometry @xcite .
both @xcite and @xcite detected optical variability and a light curve similar to that shown by polars in low states .
@xcite further showed that the amplitude of the modulations was larger at blue and ultraviolet wavelengths than in the red , consistent with a hot accretion spot responsible for the cyclotron emission .
@xcite also detected x - ray emission from sdss 1212 , conclusively proving accretion is taking place .
this paper presents cross - dispersed @xmath1 @xmath2 m spectroscopy and spectral modeling , along with @xmath7 photometry , in order to constrain the source(s ) of the observed near - infrared excess emission .
near - infrared images of sdss 1212 were acquired on 2006 june 10 at the united kingdom infrared telescope under photometric conditions ( @xmath8 ) with the wide field camera ( wfcam ; casali et al .
, in preparation ; @xcite ) .
a 5 point dither pattern was used with @xmath9 s exposures at each position ; this pattern was repeated twice at @xmath10 and thrice at @xmath6 for total exposure times of 200 and 300 s , respectively .
the data were reduced in the standard fashion .
each stack was median combined to extract a sky frame which was also normalized to generate a flat field map .
individual sky subtracted , flat fielded frames were registered and averaged for a final frame upon which to execute measurements .
the full widths at half maximum of the reduced stellar images were 2.0 and 1.9 wfcam pixels ( @xmath11 pixel@xmath12 ) at @xmath10 and @xmath6 respectively .
aperture photometry was performed with standard iraf tasks using an @xmath13 aperture and an @xmath14 sky annulus . due to the very wide field of view of wfcam ( @xmath15 ) , 15 2mass point sources in the same field as sdss 1212
were simultaneously imaged to choose amongst as possible photometric calibrators .
each 2mass star chosen as a calibrator had all of the following properties : a full width at half maximum consistent with linear accumulation ( i.e. unsaturated ) , high signal - to - noise ratios ( s / n @xmath16 100 ) in the reduced wfcam images , and reliable 2mass photometry at both wavelengths ( @xmath17 @xmath18 3.4% , @xmath19 @xmath18 6.3% ; @xcite ) .
the selected 12 calibrators are listed in table [ tbl1 ] , along with their 2mass data and their derived instrumental zeropoints .
additional photometry was acquired on 3 july 2006 at the gemini telescope north under photometric conditions with the near - infrared imager ( niri ; @xcite ) . both sdss 1212 and
the ukirt standard star fs 132 @xcite were observed in @xmath7 at nearly identical average airmasses of @xmath20 , yielding extinction corrections less than 0.003 mag .
a 5 point dither pattern was used for both science target and calibrator , with an extra frame taken for sdss 1212 .
the calibrator was observed @xmath21 s at each position while the science target was observed @xmath22 s at each position , yielding total exposure times of 50 s for fs 132 and 216 s for sdss 1212 at @xmath7 .
the data were reduced in a manner similar to the wfcam data , with separately acquired and processed flat field images .
the full widths at half maximum of the reduced stellar images were between 4.5 and 4.9 niri pixels ( @xmath23 pixel@xmath12 at f/6 ) .
aperture photometry was performed similarly using an @xmath24 aperture and an @xmath25 sky annulus .
the photometry results are listed in table [ tbl2 ] and plotted in figure [ fig1 ] . the third column in the table
gives the uncertainty in the measured aperture photometry ( due to intrinsic s / n limitations ) .
the fourth column in the table gives the calibration uncertainty : for the wfcam data , this is the standard deviation in the derived zeropoints ; for the niri data , an intrinsic 3% error was assumed .
spectroscopy was performed on 2006 june 19 and 23 at the gemini telescope south with the near - infrared spectrograph ( gnirs ; @xcite ) .
observations were executed in cross - dispersed mode with the 31.7 l / mm grating and the @xmath26 wide slit , yielding @xmath27 while effectively covering the entire near - infrared range without inter - order contamination .
data were taken at two nod positions along the @xmath28 long slit , with 300 s exposures at each nod position , repeated 6 times for a total on source integration time of 1 hr on each of two nights .
the telluric standard hip 60030 ( a7v , @xmath29 ) was observed in a similar manner with 6 s exposures at each nod position , repeated 4 times .
the raw data were processed with standard gnirs tools within the gemini iraf package version 1.9 .
these steps included flat field and pinhole map creation from calibration lamps , bad pixel correction , linearity correction , sky subtraction , image shifting and averaging , order cutting , spatial rectification , and wavelength calibration from argon lamps .
spectral extraction of each order was performed with a gnirs equivalent to the iraf apall task .
some processing of the raw data was required prior to application of the standard package tools listed above .
the first and second science frames on each night contained significant residual signal ( between 3 and 5 times larger than the read level in the unexposed portion of the array ) from the target acquisition mirror . the raw acquisition images were scaled to match the contamination level of the affected frames and subtracted off , effectively solving this problem .
additionally , there is an occasionally occurring bias level problem in one quadrant of the detector which appeared in many of the science exposures .
this noise was well removed with the nvnoise task within the gnirs tools .
both the telluric standard and sdss 1212 were processed in an identical fashion until the spectral extraction step . instead of the @xmath30 pixel diameter ( @xmath31 )
aperture used for hip 60030 , a @xmath32 pixel diameter ( @xmath33 ) aperture was used for sdss 1212 to maximize the extracted spectroscopic s / n of this faint source . flux calibration and telluric correction were performed for each order with spextool ( version 3.3 ; @xcite ) , using the general xtellcor package .
the flux levels of all four spectral orders presented here agree quite well at their boundaries and hence no inter - order adjustment was necessary .
however , the entire cross - dispersed spectrum , the average of both nights data , was adjusted slightly by a single scaling factor to best match the @xmath5 and @xmath10-band photometry in figure [ fig1 ] .
this unsmoothed spectrum is displayed in figure [ fig2 ] together with the @xmath34 points from figure [ fig1 ] .
when telescope time was requested for this study , it was not yet known that sdss 1212 was variable in any broad wavelength bands .
apart from the h@xmath0 line strength variability reported by @xcite , sdss 1212 is now known to be variable in @xmath35 , @xmath36 ( and unfiltered white optical light ) , and @xmath6 , but not in @xmath37 @xcite .
figure [ fig3 ] shows the corresponding phase of sdss 1212 for all the photometric and spectroscopic observations .
the wfcam photometry was taken during @xmath6-band minimum , by which it is meant that all the 2.2 @xmath2 m flux should be near its minimum value whether this minimum flux comes from cyclotron continuum emission , the substellar companion or a combination of sources is uncertain at present and the focus of this study .
the niri photometry , on the other hand , was taken close to the @xmath6-band maximum which occurs near orbital phases @xmath38 and 0.2 , but not at @xmath39 as for the optical and ultraviolet peak ( see 3.2 ) .
it is somewhat surprising that the nominal wfcam @xmath10 and @xmath6 magnitudes in table [ tbl2 ] are @xmath40 mag _ fainter _ than the minimum reported by @xcite , although the datasets have overlapping error bars at each passband .
as can be seen in figure [ fig3 ] , it has been verified that the wfcam observations were taken during the photometric minimum by comparing the corresponding corrected mjd with the ephemeris of @xcite .
five of the stars listed in table [ tbl1 ] were used as calibrators by @xcite .
however , the 2mass photometric errors on 3 of these objects , while mainly decent at @xmath5 , become relatively high at @xmath10 and @xmath41 ( @xmath42 @xmath18 11% , @xmath43 @xmath18 17% , @xmath19 @xmath18 25% ) , and were discarded here for this reason . in order to investigate whether this fact could account for the discrepancy between the reported @xmath7 magnitudes and more importantly , to assess the likelihood that this difference is real the instrumental zeropoints for the 3 discarded calibrators were compared with those derived with the 12 higher s / n stars used here .
those 3 stars alone yield an @xmath44 and @xmath45 which differ by 3 standard deviations ( @xmath46 mag at @xmath10 , @xmath47 mag at @xmath6 ) from the mean zeropoints used for the table [ tbl2 ] wfcam results . additionally , according to @xcite ,
not all 5 of these calibrators were present in every frame of their wirc and panic data .
the 3 stars with lowest s / n ( those discarded here ) are in fact closest on the sky to sdss 1212 , at @xmath48 , while the remaining 2 calibrators are @xmath49 distant .
hence the 3 lower quality 2mass calibrators had a good chance of being present in most , if not all , the wirc frames ( a single chip is @xmath50 ) , while the 2 higher quality calibrator stars were unlikely to be imaged in the majority of that dataset . owing to the smaller field of view of panic ( @xmath51 ) , only the 3 lowest s / n 2mass calibrators were used @xcite .
there are two additional reasons why the wfcam magnitudes in table [ tbl2 ] may be slightly discrepant from the previously published values .
first , neither study has transformed the photometry from the 2mass system into the native filter system through which the observations were taken . in the present case ,
any errors introduced by failing to transform the photometry are much smaller than the errors due to the intrinsic s / n ( represented by @xmath52 in table [ tbl2 ] ) .
also , the fact that 12 calibrators were used here would almost certainly shrink any error of this type even further due to averaging @xcite .
second , it is conceivable that sdss 1212 exhibits further variability which has so far not been documented .
the niri photometry is also listed in table [ tbl2 ] .
it has been verified that these observations were taken near the minimum hot spot viewing angle .
although this corresponds to a maximum in ultraviolet and optical light , it does not quite correspond to the cyclotron maximum as seen in the @xmath6-band light curve measured by @xcite , which occurs when the hot spot is on either limb immediately before and after the ultraviolet maximum ( @xmath39 in this paper and @xcite , @xmath53 in @xcite ) .
the spectroscopy essentially covers all orbital phases of sdss 1212 because the flux was sufficiently low ( 0.1 mjy at 1.6 @xmath2 m ) as to preclude extracting phase - resolved data the observation itself represents a benchmark near the limit of what is possible with the most advanced ground - based telescopes and instruments . the average of all data from both nights yields s / n ratios in the range @xmath54 ( average of 14 ) at @xmath10 and @xmath55 ( average of 18 ) at @xmath6 , estimated using 12 sections of 50 pixels each over the spectral ranges plotted in figure [ fig2 ] . without any inter - order adjustment , the relative levels of the three shortest wavelength orders appear to match each other well in addition to being consistent with the @xmath5 and @xmath10 photometry .
in order to perform a check on the flux level of the @xmath6-band spectroscopy , the expected photometric flux in the @xmath6-band was calculated based on the timing of the spectroscopic observations over the known @xmath6-band variability represented in figure [ fig3 ] . taking @xmath56 mag from table [ tbl2 ] and adding to that the @xmath57 mag difference between the average @xmath6-band minimum and maximum from @xcite , yields @xmath58 mag . from figure [ fig3 ] , the spectroscopic observations span , over time , an estimated average minimum to maximum ratio of 5:7 , yielding a flux - averaged magnitude of @xmath59 mag over all the exposures .
all three of these @xmath6-band fluxes are plotted in figure [ fig2 ] . as can be seen in the figure , the spectroscopic flux over the @xmath6-band is commensurate with expectations without any adjustments .
the spectrum in figure [ fig2 ] appears mostly consistent with a rayleigh - jeans slope in the @xmath5-band , where an emission line is likely detected at paschen @xmath60 and possibly also at paschen @xmath61 .
it is unclear whether any emission originates on the companion via heating by the primary , or on the white dwarf via accretion ; both are possible .
a modest excess is detected over the @xmath10-band , and is consistent with the table [ tbl2 ] photometry . at @xmath6-band
there is a definite excess throughout the window and four possible noteworthy features with varying degrees of uncertainty : a slight peak near 2.1 @xmath2 m that could be a cyclotron harmonic , brackett @xmath61 emission , the co bandhead at 2.29 @xmath2 m , as well as a significant upturn in slope near 2.3 @xmath2 m .
the near - infrared spectrum of sdss 1212 has been fitted with several spectral models and templates .
the binary components were analyzed following the method of @xcite .
a model white dwarf spectrum spanning @xmath62 @xmath2 m was generated for a non - magnetic , pure hydrogen atmosphere at @xmath63 k and log @xmath64 @xcite , using tlusty @xcite and synspec@xmath65 .
since the white dwarf has a significant magnetic field , with zeeman - split balmer lines in its optical spectrum , it is not possible at present to more precisely determine its parameters by the usual method of balmer line modeling .
the model spectrum has been normalized to the @xmath66 magnitude of the white dwarf and smoothed to match the spectral resolution of the gnirs data .
template @xmath67 @xmath2 m spectra for brown dwarfs of late l and early t type were taken from low resolution keck / nirspec datasets @xcite .
these templates were then extended to 2.5 @xmath2 m using similar near - infrared ukirt / cgs4 spectral data @xcite , in order that they span the full extent of the gnirs wavelength coverage .
the nirspec spectral templates have similar resolution ( @xmath68 ) to the gnirs data and hence no smoothing of either data was performed . however , the cgs4 data had to be interpolated between resolution elements in order to create sufficient data points to compare with the gnirs spectrum .
empirical model fluxes , determined from near - infrared photometry and parallaxes @xcite were scaled appropriately to @xmath69 pc , the nominal distance to a white dwarf with the assumed parameters listed above .
a physically realistic model for cyclotron emission was calculated utilizing the formulation and methods described in detail by @xcite and @xcite .
the model employs five critical parameters to produce the emergent cyclotron spectrum .
the magnetic field strength and electron temperature determine the cyclotron harmonic wavelengths , while the viewing angle affects the overall shape of the peaks .
also included are a parameter related to the transition from optically thick to optically thin emission ( typically well - constrained by the observations ) , and a scaling factor linked to the area of the emitting region ; a few percent of the white dwarf surface area for the models shown here .
initial attempts to fit the spectrum of sdss 1212 using cyclotron emission in a @xmath70 mg @xcite field were problematic .
the sum of the white dwarf and secondary model components does a decent job of reproducing the phase - averaged spectrum shortward of @xmath6-band .
yet the addition of such a cyclotron model component in this bandpass does not account for the upturn near 2.3 @xmath2 m in the gnirs data .
several circumbinary dust models were considered to explain this observed rise in flux at the end of the @xmath6-band , but were subsequently discarded upon the realization that this feature must be a cyclotron harmonic for the following reasons .
as pointed out by @xcite , the zero - temperature @xmath71 cyclotron harmonic for a 13 mg magnetic field lies near 2.1 @xmath2 m .
but in such a model , this cyclotron feature alone , varied by 100% , is insufficient to produce the peak - to - peak flux change ( @xmath72 mag ) seen in their @xmath6-band light curve .
once the contributions of the white dwarf and secondary are accounted for , _ all _ the remaining flux in the @xmath6-band must be 100% variable in order to explain the observed light curve ; leaving cyclotron as the only viable candidate for this extra emission .
furthermore , a 13 mg magnetic field should produce cyclotron harmonics and photometric variability within the @xmath10-band ( @xmath73 would be near 1.6 @xmath2 m ) , as seen in ef eridani and am herculis @xcite . the resolution to this dilemma
is that the magnetic field of sdss 1212 must be less than 13 mg , such that more cyclotron emission is present within the @xmath6-band , while the harmonics at shorter wavelengths are diminished .
a magnetic field of 7 mg should have two zero - temperature cyclotron harmonics within the @xmath6-band , near 2.2 and 2.5 @xmath2 m for @xmath74 and 7 respectively .
although @xcite determine an equivalent dipolar field of @xmath75 mg for sdss 1212 ( via models and assumed geometry ) , the mean surface field as determined by the measured zeeman splitting of h@xmath0 and h@xmath60 is @xmath76 mg @xcite .
as shown below , this magnetic field strength is consistent with the absence of observable cyclotron harmonics in @xmath10-band , and the presence of two harmonics in the @xmath6-band which together can explain the large photometric variability seen in this filter .
figures [ fig4 ] and [ fig5 ] display the most successful model reproductions of the gnirs cross - dispersed spectrum of sdss 1212 .
the essential difference between the two figures is that an l8 dwarf template was used for the figure [ fig4 ] plots , while a t2 dwarf template was used in the plots of figure [ fig5 ] .
it should be noted that between these two figures , a minor scaling readjusment was made to the binary components in order to best reproduce the observed spectrum of sdss 1212 .
this rescaling amounted to 8% , corresponding to a 4% change in the assumed distance to the system ;
well within the 14% minimum uncertainty arising from its poorly constrained temperature and unknown surface gravity @xcite .
the upper panels in both figures represent the best fit to the spectrum with only the binary components at @xmath69 pc ; the white dwarf model plus a brown dwarf template .
already it is clear that the @xmath10-band excess is largely reproduced by the addition of the l8 dwarf , but not as well by the t2 dwarf .
however , in both upper panels there appears to be excess flux over the @xmath77 @xmath2 m range compared to the model .
it is not clear if this is the result of poor signal or calibration in order 6 of the cross - dispersed gnirs data , or a real detection .
the middle panels of both figures add a single field cyclotron emission component to the upper panel model fits . in both figures , the models are for a 7.6 mg magnetic field , a 5 kev plasma temperature , and a viewing angle of @xmath78 , with the remaining parameters varied slightly so that the peaks of the cyclotron harmonics match the phase - averaged @xmath6-band flux .
although this set of models can account for a significant portion of the @xmath6-band flux and hence go a long way towards accounting for the photometric variability in this window , they fail miserably in the region between the cyclotron peaks .
it is perhaps noteworthy that the t2 model requires a stronger cyclotron component than does the l8 model ( due to the bluer @xmath79 color of t dwarfs versus l dwarfs ) and is more readily consistent with the @xmath6-band variability .
the lower panels in both figures display cyclotron models similar to the middle panels , but with a viewing angle of @xmath80 ; a value commensurate with the determination by @xcite .
although this angle is not strongly constrained by their data , the gnirs spectra appear to demand an intermediate viewing angle , whose result is to broaden the cyclotron harmonics , effectively blurring them together for low viewing angles @xcite . in figure
[ fig4 ] , the magnetic field employed is 7.0 mg , while in figure [ fig5 ] it is 7.2 mg , both adjusted slightly to best fit the overall spectrum when combined with their respective secondary models .
both these model fits do a decent job of reproducing the entire spectrum from @xmath81 @xmath2 m , but the l8 model matches the @xmath10- and @xmath6-band data more closely than the t2 model , as in the middle panels .
the three component , composite model fits contain a sufficient number of free parameters that their success in reproducing the @xmath81 @xmath2 m spectrum of sdss 1212 may not seem surprising , but is nonetheless informative . during initial attempts to model the gnirs spectrum , secondary spectral templates of types l6 and t5
were shown to be significantly too bright and too dim , respectively , beyond 1.5 @xmath2 m when their overall @xmath5-band flux level was matched to the data .
figure [ fig4 ] is consistent with the direct spectroscopic detection of a substellar object with spectral type l8 .
while the t2 model in figure [ fig5 ] appears somewhat inconsistent with the data , it does not rule out an earlier type t dwarf , for which available templates were lacking .
hence , the companion to sdss 1212 has a likely spectral type between l8 and t1 .
uncertainty in the distance to the white dwarf allows for some leeway in the relative contribution of the binary components at near - infrared wavelengths .
while its poorly constrained effective temperature permits photometric distances both nearer or farther than @xmath69 pc , the unconstrained mass and radius of the white dwarf are likely to overpredict its distance . a typical isolated white dwarf surface gravity , or log
@xmath64 ( 0.60 @xmath82 at @xmath63 k ; @xcite ) , is more likely to be too low than too high , especially if the primary has been accreting material over gyr timescales .
this would tend to reduce the photometric distance and lead to later spectral types for the companion .
if sdss 1212 is more massive than log @xmath64 , there is a greater likelihood that the companion has an early t spectral type .
the lower panel in the l8 model , which also matches the observed @xmath6-band spectrum well , seems to indicate that the co bandhead at 2.29 @xmath2 m may have been detected , albeit just barely .
if correct , it would be further evidence that the companion is a late l dwarf , since this co feature weakens significantly in t dwarfs @xcite . the two models for the nature of mass transfer and accretion in sdss 1212 are roche lobe overflow and wind capture .
the near - infrared observations do not distinguish between these two evolutionary scenarios , but are informative for the following reason .
the confirmed presence of cyclotron radiation adds one more observational property that sdss 1212 now shares with polars ( especially those with likely substellar secondaries in persistent low states such as ef eridani ) .
the traits it shares are : an appreciable x - ray luminosity , the presence of an accretion hot spot , a variable optical light curve with a similar shape to those of known polars in low states ( i.e. with larger modulations in the blue and ultraviolet compared to the red ) , a spin - locked binary configuration , and cyclotron emission @xcite .
occam s razor would argue that sdss 1212 is a polar .
furthermore , the hot spot x - ray luminosity of sdss 1212 is @xmath83 ergs s@xmath12 , which is nearly identical to that measured for ef eridani during its 10 yr low ; namely @xmath84 ergs s@xmath12 @xcite . while the x - ray emission from many suspected , detached , wind capture systems may be too weak ( and more importantly , not periodically variable ) to originate via accretion onto the white dwarf , but rather from activity on the m dwarf
secondary , l dwarfs simply do not have sufficient x - ray luminosities , hence favoring the low state polar model for sdss 1212 @xcite .
also , there is the issue of how such a low mass object generates sufficient wind to be captured by the magnetic white dwarf . to date
, no model has emerged which can explain how an old l or t dwarf might generate a wind on the order of @xmath85 yr@xmath12 .
the gross evolution cataclysmic variables should consist of three fundamental phases : \(1 ) a short- to long - lived phase as a detached main sequence binary which subsequently evolves to shorter orbital periods through common envelope evolution as the primary leaves the main sequence and becomes a first ascent or asymptotic giant .
angular momentum is transferred , via friction , into the envelope itself , ejecting it in the process @xcite .
if the orbit is still too large for the secondary to make contact with its roche lobe , then three mechanisms can bring this about , in principle : gravitational radiation , magnetic braking , or nuclear evolution .
the timescales of these processes depend on the masses involved , but if one assumes a canonical 0.6 @xmath82 white dwarf and a 0.2 @xmath82 secondary ( corresponding to the mass distribution peak among detached , unevolved , low mass companions to white dwarfs ; @xcite ) , then both gravitational radiation and magnetic braking operate within gyr timescales for period ranges @xmath86 hr and @xmath87 hr respectively , whereas nuclear evolution will play no role within this timeframe @xcite .
\(2 ) a phase in which the shrinking roche lobe of the secondary becomes smaller than its radius , causing unstable mass transfer onto the white dwarf . in principle
this is a runaway inspiral until the mass ratio is close to one , but if the secondary has @xmath88 @xmath82 to begin with ( as it should in the vast majority of cases based on the aforementioned companion mass distribution ) , then conservation of angular momentum competes to partially counterbalance the losses due to any braking ( magnetic or tidal ) and gravitational radiation @xcite .
at some point the trend reverses outward as the orbital evolution becomes dominated by conservation of angular momentum and any mass lost from the system entirely ( e.g. nova - like outbursts , accretion - driven winds ) .
\(3 ) a final , debatable , phase consisting of a once more detached or very low accretion rate semi - detached binary comprised of two degenerate objects in an expanded orbit which persists ( now subject only to gravitational radiation ) up to 10 gyr or so , depending on the final masses and separation .
this eventual state depends on the thermal and nuclear history of the secondary but requires a degenerate secondary core in order to cease expansion permanently .
thus , a system may end its existence as a white dwarf with a close , altered substellar companion , both objects cooling gradually toward oblivion .
the ultimate fate of catacylsmic variables outlined above is uncertain for two fundamentally different reasons .
the first issue is that the outcome itself is unknown , with end predictions for secondaries including ; complete stripping , tidal disruption , persistence as semi - detached substellar donors with very low mass transfer rates , and survival as detached substellar or even planetary mass objects @xcite .
the second issue is detectability , which itself divides into two categories : searching for either erupting systems or their burnt out remnants .
so far , astronomers who study cataclysmic variables have concentrated primarily on those systems known to be eruptive in order to search for and understand possible end states , including substellar companions .
the means of identifying such low mass secondary systems relies heavily on semi - empirical and theoretical relations between observed periods and spectral type , mass ratio , effective temperature , and mass transfer rate @xcite . without models , there is no means of determining whether a given substellar companion is migrating inward or outward , i.e. approaching or leaving its period minimum . with the prior history of these interacting systems unknown
, the beginning and end states may appear identical , theoretically .
additionally , it is difficult to know with certainty what the size of the seconday star is in such systems , especially if the mass transfer rate is low consistent with quiescence or an underfilled roche lobe ( i.e. a non - catacylsmic variable ) .
if a companion does indeed have a radius smaller than its roche lobe , then strictly speaking it is detached , and the question remains : is roche lobe contact in its future or in its past ?
the recently identified low mass transfer magnetic binaries reported in @xcite are good candidates for being detached due to the combination of the inferred accretion rates , orbital periods , and extremely low or absent x - ray emission . _
but if ef eridani can turn off for nearly 10 yr , then so can other polars _ @xcite . sdss 1212 sits somewhere on the border between a typical polar and the low accretion rate systems , having respectable x - ray accretion luminosity and a relatively low inferred mass transfer rate @xcite .
if one assumes the substellar companion to sdss 1212 has a mass and effective temperature similar to wd 0137@xmath89349b ( l8v , see table [ tbl3 ] ; @xcite ) , then its unperturbed radius should be 0.09 @xmath90 for ages between @xmath91 gyr @xcite , but its roche lobe should be located roughly at 0.11 @xmath90 for a 0.60 @xmath82 primary . therefore , the companion is either : 1 ) recently deceased , far out of thermal equilibrium and inflated as expected for an object past period minimum with the implied long history of mass - radius imbalance ; or 2 ) born dead , intrinsically detached , and slamming the magnetic white dwarf with all of its ( irradiatively driven ? ) wind .
this raises similar questions about the nature of all other low accretion rate magnetic binaries . depending on the past thermal and nuclear history of the secondary , especially if the core is degenerate , and how much mass escapes the binary completely and adiabatically
, it is quite possible for the orbital period to increase significantly over the minimum into a region where no interaction ( or wind accretion only ) occurs @xcite . without any distinct spectral features or mass - radius anomalies in such detached binaries , there may be little or no means to differentiate empirically between the pre- and post- states of these innately interacting systems .
objects must first be detected as a catacylsmic variable in order to look for substellar companions ; hence , selection effects play a large role in the number of systems suspected to harbor such cool degenerate companions @xcite .
however , this overlooks the obvious question : if the end states are detached double degenerate systems , where are they ? if one restricts a search to eruptive systems , then the potential for success is limited from the beginning for reasons stated above . why not search directly for systems which are detached and which may represent final states ?
this seems to be the only way to reconcile the well - known discrepany between the number of predicted objects past period bounce and the number observed or suspected from observation @xcite .
clearly , the best place to look for these types of end states is in the infrared where cool , jupiter - sized companions easily outshine their tiny white dwarf primaries @xcite .
there are only two issues associated with a search around white dwarfs for extinct cataclysmics the white dwarf must first have been identified or previously known and one must have sufficient sensitivity to detect its cooling leftovers .
if the end state white dwarfs are very cool and far away such that @xmath92 mag ( e.g. @xmath93 k and @xmath94 pc for @xmath95 @xmath82 ) , then only the sloan digital sky survey will detect them but will be quite insensitive to any substellar companions .
while the above conditions may not be representative , given that cataclysmic variables appear to be roughly 100 times less abundant than white dwarfs @xcite , a large sample of targets would be required to perform a statistically rigorous search ; if all cataclysmic variables are in this state , a search would only yield 1 per 100 white dwarfs ( not very encouraging ) . if these close but detached systems exist in any significant numbers , they have not yet been identified from various surveys of relatively nearby white dwarfs .
spatially unresolved l type companions , corresponding to @xmath96 @xmath82 for unevolved ages of @xmath91 gyr , are rare , with only 3 known among well over @xmath97 targets @xcite . spatially unresolved t type companions , corresponding to @xmath98 @xmath82 for unevolved ages of @xmath91 gyr , are also rare , with none detected or suspected in various surveys totalling @xmath99 targets ( @xcite ; farihi , becklin , & zuckerman 2008 , in preparation ) .
it is possible that stellar evolution has yet to produce such systems ; i.e. the initial mass function for secondaries in such systems , convolved with the finite age of the galaxy , implies that the vast majority of these systems will continue in phases ( 1 ) and ( 2 ) for up to 10 gyr .
it could be that the lack of detected end states is an observational bias ; if these systems are still marginally semi - detached then the signature of the secondary star could be masked by the accretion disk in dwarf novae and by cyclotron emission in polars .
alternatively , phase ( 3 ) never occurs and the companion is cannibalized , evaporated , stripped or tidally destroyed .
this would also explain the lack of observed end states . for completeness ,
something ought to be said about the empirical frequency of zero age cataclysmic variables with a brown dwarf secondary @xcite .
since their pre- states would consist of a white dwarf with a detached substellar companion , the above arguments and evidence apply ( @xmath100% ) . additionally , the progenitors of such pre- states would be main sequence stars with brown dwarf companions in the few au range , which are also extremely rare , occuring with a well - constrained frequency , @xmath100% @xcite .
hence , there should be no doubt that cataclysmic variables form with brown dwarf secondaries quite rarely .
the same argument which applies to detached substellar companions to white dwarfs also applies to detached stellar companions to magnetic white dwarfs , although in this latter case finding them should be easier in principle @xcite .
one might suggest a sensitive search for magnetism among known lists of close but detached , and wide common proper motion companions .
large lists of these white dwarf plus red dwarf systems exist in the following references : for known radial velocity pairs see @xcite , for spatially unresolved pairs see @xcite , and for wide common proper motion pairs see @xcite .
all of these lists contain targets which are likely to be superior to white dwarfs from the sloan digital sky survey since they are brighter and , hence , more sensitive searches could be conducted .
however , it should be noted that the main observational indicators of a magnetic white dwarf , zeeman splitting and circular spectropolarimetry , are strongest in h@xmath0 and h@xmath60 .
unfortunately , in unresolved white dwarf plus red dwarf binaries ( i.e. , the potential cataclysmic variables ) these features , especially h@xmath0 , are often completely filled - in by the red dwarf
. this may be another reason for the curious lack of detected pre - polar systems to be considered along with those already discussed by @xcite .
unfortunately , the intrinsic faintness of sdss 1212 precludes several interesting avenues of follow up , such as phase - resolved near - infrared spectroscopy or mid - infrared spectroscopy .
ultraviolet and x - ray observations are currently planned to firmly distinguish between standard mass transfer through roche lobe overflow versus accreted wind .
high resolution optical spectra of sdss 1212 may reveal the presence of elements accreted from its companion .
such observations could , in principle , distinguish between roche lobe overflow mass transfer and wind accretion as well .
there may be little or no hope of directly detecting the bare substellar companion beyond the @xmath10-band spectrum presented here .
the 7 mg cyclotron models used successfully in this paper predict the fundamental harmonic wavelength should be 15.3 @xmath2 m , thus potentially swamping every instrument aboard _
spitzer _ which might otherwise be capable of directly detecting the flux of an l8 dwarf at @xmath101 pc ; namely irac @xmath102 @xmath2 m photometry , and irs 16 @xmath2 m peak - up imaging .
however , irac observations should directly test the lower magnetic field cyclotron models employed to fit the gnirs data , which predict relatively strong emission near 3.1 , 3.8 , 5.1 , and 7.7 @xmath2 m , corresponding to the zero - temperature @xmath73 , 4 , 3 , and 2 harmonic numbers in a 7.0 mg magnetic field . yet
the possible presence of circumbinary dust emitting at irac wavelengths , as seen in several polars @xcite , may complicate the interpretation of any mid - infrared photometry for sdss 1212 .
this paper benefitted greatly from the insight of an anonymous referee , to whom the authors are commensurately grateful .
the authors would like to thank the following people : b. rodgers , g. doppman , and k. labrie ( for much appreciated help in understanding and processing gnirs data ) ; a. stephens ( for niri particulars ) ; p. hirst ( for wfcam particulars ) ; s. leggett ( for a useful discussion regarding photometric calibration errors )
. m. r. burleigh is supported by a uk science and technology facilities council advanced fellowship .
spectroscopic observations for this paper were taken as part of the gemini director s discretionary time gs-2006a - dd-5 .
photometric observations were obtained as part of gemini queue program gn-2006a - q-51 and ukirt service program 1653 .
gemini observatory is operated by the association of universities for research in astronomy , inc .
, under a cooperative agreement with the nsf on behalf of the gemini partnership : the national science foundation ( united states ) , the particle physics and astronomy research council ( united kingdom ) , the national research council ( canada ) , conicyt ( chile ) , the australian research council ( australia ) , cnpq ( brazil ) and conicet ( argentina ) .
the united kingdom infrared telescope is operated by the joint astronomy centre on behalf of the uk particle physics and astronomy research council .
this publication makes use of data products from the two micron all sky survey , which is a joint project of the university of massachusetts and the infrared processing and analysis center / california institute of technology , funded by the national aeronautics and space administration and the national science foundation .
this paper has made use of the aladin sky atlas operated at cds , strasbourg , france . 1 & j12114360 + 0137020 & @xmath103 & @xmath104 & 23.032 & 22.354 & + 2 & j12114583 + 0139374 & @xmath105 & @xmath106 & 23.083 & 22.391 & + 3 & j12114736 + 0133594 & @xmath107 & @xmath108 & 23.031 & 22.426 & + 4 & j12114758 + 0137223 & @xmath109 & @xmath110 & 23.040 & 22.479 & + 5 & j12115403 + 0132049 & @xmath111 & @xmath112 & 23.069 & 22.325 & + 6 & j12115541 + 0130399 & @xmath113 & @xmath114 & 23.041 & 22.356 & + 7 & j12120408 + 0135365 & @xmath115 & @xmath116 & 23.076 & 22.282 & 1,2 + 8 & j12120582 + 0135157 & @xmath117 & @xmath118 & 22.980 & 22.404 & 1,2 + 9 & j12120662 + 0138153 & @xmath119 & @xmath120 & 23.075 & 22.405 & + 10 & j12120989 + 0135259 & @xmath121 & @xmath122 & 22.908 & 22.115 & 1,2 + 11 & j12121658 + 0137042 & @xmath123 & @xmath124 & 23.039 & 22.452 & 2 + 12 & j12121667 + 0135257 & @xmath125 & @xmath126 & 23.036 & 22.383 & 2 + 13 & j12122238 + 0130101 & @xmath127 & @xmath128 & 23.058 & 22.410 & + 14 & j12122266 + 0137425 & @xmath129 & @xmath130 & 23.011 & 22.470 & + 15 & j12122647 + 0140477 & @xmath131 & @xmath132 & 23.070 & 22.412 & + | the nature of the excess near - infrared emission associated with the magnetic white dwarf commonly known as sdss 1212 is investigated primarily through spectroscopy , and also via photometry .
the inferred low mass secondary in this system has been previously detected by the emission and variation of h@xmath0 , and the @xmath1 @xmath2 m spectral data presented here are consistent with the presence of a late l or early t dwarf . the excess flux seen beyond 1.5 @xmath2 m in the phase - averaged spectrum
is adequately modeled with an l8 dwarf substellar companion and cyclotron emission in a 7 mg magnetic field .
this interesting system manifests several observational properties typical of polars , and is most likely an old interacting binary with a magnetic white dwarf and a substellar donor in an extended low state . |
let @xmath1 be a field of characteristic other than @xmath2 or @xmath3 .
a picard curve @xmath4 is a smooth , projective , absolutely irreducible cyclic trigonal curve of genus @xmath3 . over any extension @xmath5 possessing a primitive cube root of unity , the required degree @xmath3 morphism @xmath6 is galois , and one always may recover from this map a smooth affine model of the form @xmath7 where @xmath8 $ ] is a binary quartic form .
picard curves are always non - hyperelliptic , and provide the simplest examples of curves of gonality @xmath3 .
they have generated interest from geometric , arithmetic , and computational perspectives .
picard curves and their moduli arise in connection with some hilbert problems @xcite , and the monodromy of such moduli spaces have been studied in @xcite .
efficient algorithms have been found for arithmetic on the associated jacobian varieties over finite fields @xcite .
algorithms for counting points @xcite and statistics on point counts for picard curves and the associated jacobian varieties have also been studied @xcite . in the present article ,
we enumerate all @xmath0-isomorphism classes of picard curves @xmath9 with good reduction away from @xmath3 . because both the geometry and the arithmetic of such curves is exceptional
, they have some applications to an open question of ihara , which we discuss briefly at the end of the paper . in @xcite ,
n. p. smart enumerated all curves @xmath9 , up to @xmath0-isomorphism , subject to the following constraints : * @xmath10 has genus @xmath2 , * @xmath10 has good reduction away from the prime @xmath2 .
because all such curves are hyperelliptic , they are naturally equipped with a degree @xmath2 morphism to @xmath11 . to carry out the exhaustive search , smart
first establishes a one - to - one correspondence @xmath12$-equivalence classes } \\
\text{of binary sextic forms } \\
\text{good away from $ 2 $ } \end{array } \right\}.\ ] ] ( the definition of @xmath13$]-equivalence is explained in [ sec : notationsbackground ] . ) the correspondence is straightforward : given a representative @xmath14 of a class from the set on the right , the associated curve @xmath10 has an affine model of the form @xmath15 .
however , this association is not quite one - to - one ; if @xmath16 is a quadratic extension unramified away from @xmath2 , then the set on the right can not distinguish between a curve @xmath10 and its quadratic twist by @xmath17 .
the subscript `` twists '' indicates that the association is a bijection up to this family of twists ; of course , recovering the distinct twists is trivial once one curve in the family is known , since there are only a finite number of extensions @xmath16 with the necessary properties .
any class from the right hand set has a representative @xmath14 whose numerical invariants may be expressed in terms of solutions to an @xmath18-unit equation over the splitting field of @xmath19 . by (
* lemma 3 ) , this field is bounded in degree and may ramify only at @xmath2 .
hence , there are only finitely many fields to consider , and each yields only finitely many solutions to the @xmath18-unit equation .
smart then gives the results of a complete search ; first finding all possible @xmath18-unit equation solutions ( thus all possible numerical invariants ) , and then determining at least one representative for each of the desired classes of sextic forms . recovering the genus @xmath2 curves
is now immediate from the correspondence .
the process does not boil down to brute force , however .
smart employs several nontrivial algorithms to reduce the search to one of manageable size . as a hyperelliptic curve @xmath20
is equipped with a degree @xmath2 morphism @xmath21 , one might suspect that the generalization of smart s approach to the case of picard curves will be immediate .
that is , should one expect the association @xmath22$-equivalence classes } \\
\text{of binary quartic forms } \\ \text{good away from $ 3 $ } \end{array } \right\}\ ] ] also to be a one - to - one correspondence , up to twists ?
unfortunately , no : the association of classes of forms to isomorphism classes of picard curves is not even well - defined , as there are @xmath23$]-equivalent quartic forms @xmath19 , @xmath24 whose associated picard curves @xmath25 are not isomorphic over @xmath0 . in [ sec : picardcurves ] , we introduce a finer notion of equivalence which will let us establish a correspondence between isomorphism classes of picard curves and equivalence classes of forms .
once this correspondence is established , the search for picard curves with good reduction away from @xmath3 follows smart s program in many important respects .
however , the project diverged from smart s work in the more delicate nature of the curves involved , which necessitated developing modified algorithms for determining all equivalence classes .
this paper also includes one theoretical improvement to smart s approach : lemma [ lemma : infinite place ] makes the use of a @xmath26-adic lll reduction unnecessary .
we may avoid using the bounds on @xmath26-adic logarithmic forms established by yu @xcite , as well as the @xmath26-adic lll algorithm used by smart @xcite .
the primary result of the paper is the following : up to @xmath0-isomorphism , there are exactly @xmath27 picard curves @xmath9 with good reduction away from @xmath3 . moreover , up to cubic twists , there are @xmath28 classes of such curves ; each of these contains exactly @xmath3 @xmath0-isomorphism classes of picard curves ( corresponding to the twists by @xmath29 ) . in [ sec :
notationsbackground ] , we introduce some notation and review some background on binary forms from smart s methods . in [ sec : picardcurves ] , we introduce a finer notion of equivalence and establish the desired one - to - one correspondence between classes of picard curves and classes of objects called quintic - linear pairs . in [ sec : quarticsequiv ] and [ sec : quartics0equiv ] , we explain how to recover a representative from each equivalence class of quintic - linear pairs over @xmath0 with good reduction away from @xmath3 . in [ sec : solvingsunit ] , we describe the methods used to solve the @xmath18-unit equation . in [ sec : conclusions ] , we provide explicit models for the classes of picard curves , and discuss further directions and applications of the work .
the authors are delighted to acknowledge the contribution of michel brner , at the institute of pure mathematics , ulm university .
he brought to our attention that the list of isomorphism classes of curves provided in an earlier version of this paper was incomplete !
brner , after reading our previous version , had implemented this algorithm independently , and contacted us when his own computation yielded classes we had missed . upon review of our own implementation
, we discovered an error that had led to the omissions .
we are extremely grateful to him for the assistance .
we emphasize that this computation has now been carried out independently by both brner and the authors , and the results are in agreement .
in addition , the authors are grateful to jeff achter , wai kiu chan , norman danner , david grant , rachel pries , and john voight for many helpful conversations during this project .
we appreciate the detailed comments of the anonymous referees , which certainly improved the clarity of the paper and alerted the authors to an error in a previous version .
this research was partially supported through wesleyan university s van vleck fund .
let @xmath30 be a number field with @xmath31 real embeddings and @xmath32 pairs of conjugate complex embeddings ( so @xmath33 = n_1 + 2n_2 $ ] ) , and let @xmath34 be a finite set of places of @xmath17 which includes all the infinite places of @xmath17 and exactly @xmath35 finite places . by convention , we enumerate first the finite places , then the real , then the complex infinite places .
let @xmath36 be the ring of integers of @xmath17 , and let @xmath37 be the ring of @xmath18-integers in @xmath17 .
if @xmath38 and @xmath39 , then the ring @xmath37 has the form @xmath40.\ ] ] for any @xmath37-ideal @xmath41 , let @xmath42 denote an @xmath36-ideal , coprime to the finite places of @xmath18 , such that @xmath43 .
the _ @xmath37-norm _ of @xmath41 is @xmath44}.\ ] ] for any @xmath45 , we set @xmath46 .
we let @xmath47 be the group of @xmath18-units in @xmath17 .
let @xmath48 be a fixed , finite set of rational primes , together with @xmath49 , the infinite place of @xmath0 .
often the set @xmath18 will contain precisely those places of @xmath17 lying above a place of @xmath48 .
even more explicitly , we usually have in mind @xmath50 .
we record the following fact for later use in solving @xmath18-unit equations .
there are only finitely many fields @xmath17 which satisfy the hypotheses of the lemma ; a complete list may be found from the number field database of jones and roberts @xcite .
the verification of ( i ) and ( ii ) is immediate for each field by explicit computation .
table [ table : nf_disc_3 ] records all the number fields for which the lemma applies , and which either have degree @xmath55 , or are the galois closure of such a field .
in particular , fields @xmath56 are galois ; @xmath57 , @xmath58 , @xmath59 are conjugate , and @xmath60 is the galois closure of @xmath57 .
( there are @xmath2 additional sextic fields unramified away from @xmath51 , but we will not need these in the sequel . ) let @xmath62 be a rational prime number , and let @xmath17 be a number field .
we let @xmath63 denote the maximal @xmath62-cyclotomic extension of @xmath17 , and let @xmath64 denote the maximal pro-@xmath62 extension of @xmath63 unramified outside @xmath62 .
an abelian variety @xmath65 is said to be _ heavenly _ at @xmath62 if the field @xmath66)$ ] is a subfield of @xmath67 . and
@xmath68 are pronounced _ ten _ and _ san _ , respectively . ]
let @xmath69 denote the absolute galois group @xmath70 , and let @xmath71 , the projective line ( over @xmath0 ) with three points deleted .
there is a canonical outer galois representation attached to the pro-@xmath62 algebraic fundamental group of @xmath72 , @xmath73 we let @xmath74 denote the fixed field of the kernel of @xmath75 .
it is well known that @xmath76 , and ihara has asked , specifically in the case @xmath38 , whether or not the two fields coincide .
to rephrase the second condition , there exists , over some field extension , a morphism @xmath78 branched only over @xmath79 , and whose galois closure is of @xmath62-power degree .
let @xmath80 denote the jacobian variety of @xmath10 .
then anderson and ihara have given a necessary condition on the morphism @xmath81 ( * ? ? ?
. 3.8.1 ) which implies @xmath82 ) \subseteq \san$ ] .
the picard curves enumerated in the present article all have good reduction away from @xmath3 , and naturally admit a degree @xmath3 morphism @xmath83 .
thus , in the spirit of @xcite or @xcite , we may use the criterion of anderson and ihara if there exists a morphism @xmath84 such that @xmath85 satisfies the conditions ( i1 ) and ( i2 ) above .
for example , on each of the curves @xmath86 there exists a degree @xmath87 morphism @xmath78 defined in affine coordinates by @xmath88 it is necessarily galois over @xmath89 , as it corresponds to a composition of degree @xmath3 kummer extensions : @xmath90 .
the verification of anderson and ihara s criteria is immediate , and so we obtain the following .
let @xmath10 be one of the nine picard curves over @xmath0 with good reduction away from @xmath3 , possessing an affine integral model of the form .
let @xmath80 denote the jacobian of @xmath10
. then @xmath91 ) \subseteq \san({\mathbb{q } } , 3)$ ] .
a. peth and b. m. m. de weger .
products of prime powers in binary recurrence sequences . i. the hyperbolic case , with an application to the generalized ramanujan - nagell equation .
, 47(176):713727 , 1986 . c. rasmussen .
an abelian surface with constrained 3-power torsion . in _ galois - teichmller theory and arithmetic geometry _ , volume 63 of _ adv .
pure math .
_ , pages 449456 .
japan , tokyo , 2012 . | inspired by methods of n. p. smart , we describe an algorithm to determine all picard curves over @xmath0 with good reduction away from 3 , up to @xmath0-isomorphism . a correspondence between the isomorphism classes of such curves and certain quintic binary forms possessing a rational linear factor is established
. an exhaustive list of integral models is determined , and an application to a question of ihara is discussed . |
in this paper we consider the effects of the spontaneous breaking of time - reversal ( @xmath5 ) invariance in electronic systems . this is a problem of considerable current interest particularly in the context of strongly correlated systems . while the physics of strong correlation is important , many aspects of spontaneous time - reversal symmetry breaking are not well understood even at the level of weakly coupled systems , well described by fermi - liquid theory .
the problem that we will consider is that of the possible quantum phase transitions to states in which time - reversal invariance is spontaneously broken in electronic systems with several fermi surfaces , and how to classify them .
one of the most important consequences of the spontaneous breaking of time reversal is that these ground states may exhibit a spontaneous ( nonquantized ) anomalous hall effect .
more specifically , we consider fermi systems with multiple fermi surfaces with condensates in the particle - hole channel that break time - reversal invariance . as a consequence ,
these systems have a nontrivial _ relative _ fermi - surface berry curvature which quantifies the strength of the time - reversal symmetry breaking .
the theory that we present here has a close connection with haldane s analysis of the anomalous hall effect as a berry curvature on the fermi surface .
@xcite the nontrivial new effect that results from these states is that they exhibit a spontaneous anomalous hall effect , _
i.e. _ , present even in the absence of extrinsic effects such as magnetic impurities or external magnetic fields .
time - reversal symmetry breaking in the absence of external magnetic fields or spontaneous spin ordering has been a focus of interest in condensed - matter physics for quite some time , at least since the discovery of high @xmath6 superconductivity in the copper oxide materials .
quite early on it was postulated that frustrated two - dimensional ( 2d ) quantum antiferromagnets may have `` chiral spin liquid '' phases ( or ground states ) , translationally invariant states without magnetic long - range order.@xcite the known behavior of high @xmath6 superconductors does not appear to be consistent with a spin liquid ground state . the discovery of time - reversal symmetry - breaking effects in recent experiments on and in underdoped ( and similar systems ) has renewed the interest in understanding time - reversal symmetry - breaking phases in strongly correlated electronic systems . the main purpose of this paper is to inquire if it is in principle possible to have an electronic system with a nonmagnetic translationally invariant ground state that breaks spontaneously time - reversal invariance .
such a system would exhibit a spontaneous hall effect and a kerr effect even in the absence of disorder and/or magnetic fields .
we will focus here on simple models of quantum phase transitions in electronic systems , without ferromagnetism or any other form of long - range magnetic order .
we will further assume that these systems have well - defined electronic quasiparticle excitations and are hence extensions of fermi - liquid theory .
this assumption is valid for any mean - field approaches and as will be shown in sec .
[ sec : fluctuations ] , is valid even if the fluctuations of the order parameter around its mean - field value are considered for lattice models .
we will see that , even in this `` weak - coupling '' approach , states with the desired properties are physically sensible .
( naturally , the naive applicability of the details of this theory to a regime of strong correlations , necessary in the context of the cuprates , is questionable . ) within this weak - coupling approach , time - reversal breaking phases can be described in terms of properties of the resulting one - particle states and of their effective fermi surfaces . of particular importance
is the fact that a system with @xmath2 fermi surfaces obtains an `` emergent '' ( and approximate ) local _ gauge _ ( in momentum space ) @xmath7 symmetry near the fermi - liquid fixed point . in sec .
[ sec : gauge ] we show that the natural way to represent this structure is in terms of an abelian gauge theory of @xmath2 gauge fields .
these gauge fields describe the quasiparticle berry phases @xcite in the sense discussed recently by haldane .
@xcite for systems with one or two band(s ) , the structure of @xmath5 symmetry breaking is described by these gauge fields and the symmetry properties of the quasiparticle dispersion relation under space inversion . however , for more than two bands , more complex @xmath5 symmetry - breaking phases arise involving additional time - reversal breaking operators which are neither berry connections nor the space inversion symmetry of the quasiparticle dispersion relations .
nevertheless , the berry connections and the inversion symmetry of the quasiparticle dispersion relation still describe a large class of the @xmath5 symmetry - breaking states even in multiband models
. we will only consider the @xmath5 symmetry - breaking phases described by these gauge fields ( berry phases ) and the inversion symmetry of quasiparticle dispersion relations . within this constraint ,
the systems that we describe are invariant under the combined transformation of @xmath8 , where @xmath9 is the chiral transformation ( _ i.e. _ , a mirror reflection ) and @xmath10 stands for space inversion . with the @xmath8 symmetry
, the @xmath5 symmetry - breaking states can be classified into two classes , according to the accompanying @xmath9 or @xmath10 symmetry breaking .
we refer to the states that preserve all three symmetries @xmath9 , @xmath10 , and @xmath5 , as the normal states . as for the @xmath5 symmetry - breaking states ,
if the inversion @xmath10 symmetry is also broken but the chiral @xmath9 and the combined @xmath11 symmetries are preserved , these states will be referred to as type @xmath3 .
in contrast , the states that break @xmath5 and @xmath9 but preserve @xmath12 and @xmath10 will be referred as the type @xmath4 states .
the states that break all three of @xmath9 , @xmath10 , and @xmath5 are considered as a mixing of types @xmath3 and @xmath4 .
obviously , type @xmath3 states have no kerr or hall effect , but type @xmath4 may have .
the type @xmath3 state is somehow trivial if we notice that the momentum @xmath13 changes sign under @xmath10 or @xmath5 .
hence , in this paper , we mostly concentrate on the type @xmath4 states . using the berry connections , we show that type @xmath3 phases may appear in one - band or multiband models .
however , the type @xmath4 phases can only be found in multiband models .
the phase transitions from the normal phase to the type @xmath4 phase can be classified into two different scenarios depending on whether the band structure has degeneracy points or not ( degeneracy lines or areas usually require fine tuning and will not be considered ) . after exploring the general theory ,
we use mean - field theory to investigate the @xmath5 symmetry breaking in several specific models . in sec . [
sec : order_parameter ] , we study the general symmetry properties of the ( fermion bilinear ) order parameters in the particle - hole channel for systems whose band structure contains no degeneracy point and show that the type @xmath4 @xmath5 symmetry - breaking states requires two order parameters . in sec .
[ sec : isotropic ] , we present a mean - field study of 2d fermi liquids with continuous rotational and translational symmetries , and time - reversal invariance . in this section
we discuss the possible patterns of spontaneous breaking of time - reversal invariance , inversion and chiral symmetries , and rotational invariance in interacting metallic systems .
although the models we discuss here use the framework of the landau theory of the fermi liquid , the patterns of symmetry breaking that are found , as well as the resulting phenomenology , are of more general interest . in a landau - type model with four - fermion forward scattering interactions , analogous to the type discussed in refs . @xcite and @xcite , time - reversal symmetry - breaking phases can be stabilized and are usually accompanied by rotational symmetry breaking . in a one - band model , the type @xmath3 phases can be reached through a pomeranchuk instability @xcite in _ odd angular momentum channels _ ( with angular momentum @xmath14 ) , nematic - like phases with broken space inversion and time reversal .
however for two - band models type @xmath4 phases may also appear , and have a similar structure to the @xmath15 phases in fermionic systems with spin described in refs .
[ ] and [ ] . in this section
we construct the phase diagram . here
we also evaluate the hall conductance for these phases , which is not quantized since these phases are gapless and conducting .
we also show that the hall conductance found here is related with a topological index , the kronecker index of the homotopy mappings @xmath16 , _ i.e. _ , @xmath17 .
this in turn implies that the @xmath5 symmetry breaking in these phases is stable against adiabatic perturbations , even though the actual value of the _ unquantized _ hall conductance is not universal and can be changed continuously . in sec .
[ sec : lattice_model ] , we generalize these @xmath5 symmetry - breaking phases to lattice models and discuss subtle effects arising from the degeneracy points in the band structures . in particular we show that without degeneracy points , the time reversal @xmath5 and chiral @xmath9 symmetry - breaking phase can be reached from a normal fermi liquid either by a direct first - order transition or by two separate phase transitions through an intermediate phase characterized by rotational symmetry breaking .
but in the presence of degeneracy points , the direct transition between the @xmath5 and @xmath9 symmetry - breaking phase and the normal fermi liquid may be second order . finally , in sec .
[ sec : discussion ] , we present a discussion of the experimental consequences of this work .
the relation between this work and its particle - particle channel counterpart is also discussed , as well as the similarities and differences with the phases studied in ref .
we present details of the calculations for two - band models in appendix [ app : two_band ] .
the topological and physical meaning of the wilson loops introduced in sec .
[ sec : gauge ] is discussed in appendix [ app : wilson_loop ] . in appendix
[ app : hall_cond ] , we present the details of the calculation of the hall conductivity , and in appendix [ app : z2 ] , the symmetry of the @xmath18 and @xmath19 phases ( to be defined below ) is discussed .
in this section , we study the general properties of the spontaneous @xmath5 symmetry breaking for a fermionic system , which we will assume to be well described by an effective fermi liquid , _ i.e. _ , a fermionic system with well - defined quasiparticle excitations which are asymptotically free at low energies .
we will also assume that time - reversal invariance is not broken explicitly and hence that there is neither an external magnetic field nor any sort of magnetic long - range order .
we will consider systems without magnetic impurities , trapped magnetic fluxes , or other explicit extraneous time - reversal symmetry - breaking effects . for one- and two - band models
, we will show that : 1 .
the @xmath5 symmetry - breaking effects are represented either by the existence of berry phases or by the symmetry properties of the quasiparticle dispersion relation under space inversion ; 2 .
there is a @xmath20 symmetry , where @xmath5 is time - reversal , @xmath9 is a chiral transformations ( reflection across a suitable mirror plane ) , and @xmath10 is space inversion ; 3 . in the absence of explicit breaking of @xmath5 ,
the total berry phase of all the bands is zero , @xmath21 ; 4 .
there is no type @xmath4 @xmath5 symmetry breaking in a one - band model ; 5 .
degeneracy points of the effective band structure ( defined later ) have an associated quantized berry flux @xmath22 , with integer @xmath23 ; 6 .
systems with and without degeneracy points have different properties when undergoing a phase transition to a type @xmath4 @xmath5 symmetry - breaking phase . here
type @xmath3 ( breaking @xmath10 and @xmath5 ) and type @xmath4 ( breaking @xmath9 and @xmath5 ) refer to the two types of @xmath5 breaking phases discussed in the introduction . for models with more than two bands , we will show that by assuming @xmath24 , all other conclusions above can be generalized easily .
the low - energy properties of a fermi liquid are described by its spectrum of quasiparticle excitations , _ i.e. _ , bloch waves and their dispersion relation .
the dispersion relation @xmath25 , where @xmath23 is the band index with @xmath26 for an @xmath2 band model , transforms to @xmath27 under time reversal @xmath5 or space inversion @xmath10 but is invariant under chirality @xmath9 or the simultaneous action of @xmath5 and @xmath10 . hence
, the odd part of the dispersion relation , @xmath28 , describes type @xmath3 @xmath5 symmetry breaking .
the bloch waves may also contain information of the @xmath5 symmetry breaking .
due to the ( perturbative ) irrelevance at low energies of the quasiparticle interactions under the renormalization group ( rg ) , @xcite a fermi liquid with @xmath2 bands is invariant under a @xmath1 gauge transformation @xmath29 where @xmath30 is the bloch wave function of the band @xmath23 .
this @xmath1 gauge symmetry , associated with independent redefinitions ( gauge transformations ) of the phase of the quasiparticle bloch states for each band at each wave vector @xmath13 , is an `` emergent '' symmetry , asymptotically accurate only close enough to the fermi - liquid fixed point .
away from this fixed point , the irrelevant quasiparticle scattering processes make this @xmath1 symmetry an approximate one .
this effect can be studied perturbatively as will be shown in sec .
[ sub : sec : other_scatterings ] , but can ( and will ) be ignored for the purposes of the present discussion . to remove the redundant degrees of freedom , one defines the overlap matrix @xcite @xmath31 the diagonal terms , @xmath32 , are berry connections which under the gauge transformation of eq .
transform as gauge fields : @xmath33 the off - diagonal terms , for @xmath34 , transform instead as @xmath35 which can not be regarded as gauge fields .
clearly the overlap matrix @xmath36 are the matrix elements of the position operator in bloch states .
@xcite the diagonal terms are directly related with the `` anomalous velocity '' in the semiclassical theories of bloch waves .
@xcite it is well known from the theory of the hall effect ( see , for instance , haldane s work@xcite and references therein ) that an external magnetic field induces a nontrivial berry curvature .
this affects all the bands in essentially the same way . in what follows
we will assume that time - reversal invariance is not broken explicitly by external fields .
thus , the total berry curvature vanishes as required by eq . .
we now restrict our discussion on @xmath5 symmetry breaking that can be described by the diagonal terms @xmath37 alone .
as shown in appendix [ app : two_band ] , this assumption is automatically satisfied for all one- and two - band models . for systems with three or more bands it is also possible to have phases that break time - reversal invariance which are described purely by off - diagonal operators , @xmath38 ( with @xmath34 ) . even though they do break time - reversal , these states do not have a berry phase and , consequently , will not have a spontaneous anomalous hall effect .
we will discuss in sec .
[ sec : lattice_model ] that one of the so - called varma loop states , @xmath39 in the notation of ref .
[ ] , is an example of a time - reversal symmetry - breaking state in a three - band model without an anomalous hall effect . from now on , only the diagonal terms @xmath37 will be considered .
their , gauge - invariant , physical degrees of freedom are described by the wilson loops @xmath40 which is also the berry phase .
@xcite notice that in principle one can consider any path @xmath41 in momentum space . in practice , for a gapless system the path of physical interest coincides with the location of the fermi surfaces of the bands ( see below ) . under the action of the chiral transformation @xmath9 , space inversion @xmath10 , and time reversal @xmath5 , the wilson loops @xmath42
transform as @xmath43 the chiral transformation @xmath9 reverses the orientation of the path , @xmath44 , and hence , it transforms @xmath42 into its complex conjugate .
the space inversion operator @xmath10 changes momentum @xmath13 to @xmath45 , which changes the integration contour but preserves its direction .
the time - reversal operator @xmath5 is antiunitary .
it changes @xmath13 to @xmath45 as @xmath10 but also changes the wilson loop to its complex conjugate , which is also how @xmath9 acts .
therefore , @xmath42 is invariant under @xmath8 . since the dispersion relation is invariant under @xmath11 ( and @xmath9 ) ,
then the system must be invariant under @xmath8 .
this is one of the key results of this paper , which makes the classification of the types @xmath3 and @xmath4 @xmath5 symmetry - breaking states possible . as we noted above , in the absence of magnetic fields ( and of any other explicit breaking of time reversal )
, there is a constraint over the total berry phase , @xmath46 which implies that there is no berry phase associated with the charge sector , the overall @xmath0 gauge group .
however , in the presence of magnetic fields , the phase of the bloch wave can not be determined in a unique and smooth way over the entire brillouin zone .
@xcite this invalidates the assumptions behind eq .
, leading to a nontrivial berry phase in the charge @xmath0 sector and a nonvanishing hall conductance . on the other hand , in the absence of external magnetic fields , although the constraint of eq . prevents the charge @xmath0 sector to obtain a berry phase , a nontrivial _ relative berry phase _ between different bands is still allowed .
this is the key point in our study of a spontaneous @xmath5 symmetry breaking without magnetic ordering that we are interested in here . for a one - band model , the constraint that the total berry phase must be trivial , @xmath47
, implies that the wilson loops must be real , @xmath48 .
hence , the wilson loop is an eigenvector of the chiral transformation @xmath9 , _
i.e. _ , @xmath49 , and the @xmath11 symmetry always holds .
thus , for one - band models , only type @xmath3 time - reversal spontaneous symmetry breaking is allowed .
let us consider now the case of two - band models which allow for a richer structure . in this case eq .
now becomes @xmath50 .
hence , only one berry phase @xmath51 ( or equivalently @xmath52 ) is linearly independent . for a state with @xmath9 , or equivalently @xmath11 , symmetry ( the normal state or the type @xmath3 state ) ,
again one obtains @xmath53 , which quantizes the berry phase to be @xmath54 , with integer @xmath23 .
conversely , if @xmath9 _ and _ @xmath11 are broken , the symmetry breaking is type @xmath4 . in this case , @xmath5 symmetry breaking in two - band models can be described by @xmath55 obtaining an imaginary part .
equivalently , the berry phase @xmath51 becomes nonquantized for a type @xmath4 time - reversal symmetry breaking .
the effective one - particle hamiltonians of a two - band model has the form of a @xmath56 hermitian matrix whose coefficients are smooth functions of the momentum @xmath13 . in appendix
[ app : wilson_loop ] we use standard arguments@xcite to relate the berry phase for this system to a wess - zumino term , familiar from the path integral for spin ( see , _
e.g. _ , ref . [ ] ) .
the berry phase actually equals half of the wess - zumino term . for a specific contour , where @xmath41 coincides with the fermi surface , the berry phase is proportional to the hall conductance , @xcite as reviewed in appendix [ app : hall_cond ] . for an insulator , which does not have a fermi surface ,
the contour is the boundary of the brillouin zone .
this leads to the well - known quantization of the wess - zumino term as a chern number , and it implies the quantization of the hall conductance .
however , for a metallic state , the contour @xmath41 is the fermi surface , and the manifold is no longer compact . as emphasized by haldane,@xcite in this case
the hall conductance is in general not quantized , which leads to anomalous hall effect . in sec .
[ sub : sec : hall_cond ] , the anomalous hall conductance for the metallic spontaneous type ii @xmath5 symmetry - breaking phase in a specific model will be computed within a mean - field approximation .
we end the discussion of this section by emphasizing one property of the berry phase for later use . for systems with @xmath11 symmetry , _ i.e. _ , @xmath57 , field strength @xmath58 is zero away from degeneracy points ( _ i.e. _ , for points in momentum space where the bands have different energy ) .
however , at degeneracy points ( _ i.e. _ , points in momentum space where two or more bands have the same energy ) the field strength may have berry flux @xmath59 , with integer @xmath23 .
as will be shown later , this difference of the berry phase leads to different phase transitions in type @xmath4 @xmath5 symmetry breaking . in the rest part of this paper
, we will study specific microscopic models where the conclusions above are applied .
in this section , we study systems without degeneracy point in the band structure by writing down order parameters in the particle - hole channel that preserve both translational and charge @xmath0 symmetries , the most general ground - state expectation values of bilinears in fermion operators of the form @xmath60 where @xmath61 and @xmath62 are the fermionic creation and annihilation operators and @xmath13 is the momentum of the quasiparticle ; the indices @xmath23 and @xmath63 label the bands . the spin indices are dropped since we are not considering spin ordering .
each hermitian matrix @xmath64 defines a real order parameter @xmath65 ( complex order parameters are two real order parameters ) . in the absence of band crossings ,
the @xmath11 symmetry implies the existence of a special gauge in which @xmath66 . with this gauge choice , under @xmath10 and @xmath5 ,
@xmath67 transforms as @xmath68 the fermion bilinears defined in eq .
can always be expressed in such a way that transform irreducibly under the symmetries of the system .
here we are interested in particular in their transformation properties under time reversal . in mean - field theories , such as the one we will discuss in sec .
[ sec : isotropic ] , the one - particle effective hamiltonian depends linearly on these fermionic bilinear order - parameter fields .
thus , a non - vanishing expectation value of the order parameter breaks the symmetry .
it is not possible to have a state that breaks spontaneously time - reversal invariance in a way that can not be compensated by another symmetry transformation . in sec .
[ sec : single ] we will construct a state in one - band model with a ground state that breaks @xmath5 which however must also break @xmath10 , space inversion .
this model , which has an order parameter in the particle - hole channel with angular momentum @xmath69 , is an explicit representation of the `` varma loop '' state ( @xmath70 ) discussed recently by varma as a @xmath5 breaking state@xcite .
the state is , however , also odd under @xmath10 but invariant under @xmath71 and under chirality @xmath9 .
thus , this is a type @xmath3 state in the classification discussed in the introduction . as a consequence
this state does not have a spontaneous anomalous hall effect or a uniform kerr effect in the absence of external magnetic fields or defects .
( in this sense this pattern of time - reversal symmetry breaking is analogous to that of a nel antiferromagnet , in which time reversal and translation invariance by one lattice constant are broken but the combination of both is not . ) in order to obtain a state with time - reversal symmetry breaking but without space inversion symmetry breaking , it is necessary to have at least two bands . for two - band models ,
even though the natural symmetry in this case is @xmath72 , it can be naturally embedded in a the larger @xmath73 group ( even though it is not a symmetry ) . to obtain a ground state with broken time - reversal invariance
it is necessary to break @xmath73 completely ( down to its center , the @xmath74 subgroup ) .
this requires that two ( non - commuting ) generators of @xmath73 must be broken in the ground state .
thus the order parameter has two components and can not be made real by a gauge transformation .
similarly , the eigenstates of the effective one - particle hamiltonian are complex .
a system with these properties will have a nontrivial berry connection .
it turns out that it also breaks chirality , @xmath9 , but it is invariant under @xmath75 .
hence , this state corresponds to a type @xmath4 time - reversal symmetry breaking .
a state of this type has a spontaneous anomalous hall effect and a kerr effect even in the absence of external magnetic fields and defects .
examples of states of this type in the particle - particle channel are the well - known @xmath5 breaking spin triplet @xmath76 and spin singlet @xmath77 superconducting condensates .
similarly , the @xmath77 @xmath78-density wave ( @xmath78dw ) state breaks time reversal and translation invariance [ with @xmath79 in the particle - hole channel . in the following sections we present some simple models containing the types @xmath3 and @xmath4 @xmath5 symmetry - breaking states .
for simplicity , only one- and two - band models will be studied .
however , the conclusion can be generalized to multiband models with little effort .
let us consider a 2d isotropic fermionic system with hamiltonian @xmath80 where @xmath25 is the single - particle kinetic energy for band @xmath23 and @xmath81 is the chemical potential .
@xmath82 is the interacting part of the hamiltonian .
here we only consider the forward - scattering interactions . in a one - band model
, @xmath82 can be expanded into different angular momentum channels , denoted by the angular momentum quantum number @xmath83 .
if the coupling constant in some channel , @xmath84 , is attractive and strong enough to violate the pomeranchuk condition , @xmath85 , where @xmath86 is the density of states at the fermi level , then the spherical fermi surface becomes unstable , leading to a spontaneously distorted fermi surface as shown in fig .
[ fig : pomeranchuk ] .
this is the pomeranchuk instability .
@xcite the case of the @xmath87 channel is the electron nematic phase .
@xcite similar condensates for other even angular momentum channels can also be ( and have been ) considered . in the case of odd angular momentum channels
both time reversal @xmath5 and space inversion @xmath10 are spontaneously broken , but the product @xmath88 remains unbroken as these states are not chiral : @xmath9 is unbroken .
these states correspond to the type @xmath3 @xmath5 symmetry - breaking states . as can be seen from figs .
[ fig : pomeranchuk](b ) and ( d ) , the chiral symmetry is preserved and the @xmath5 symmetry can be recovered by space inversion .
let us now consider the case of a two - band model .
it can be either a system with two bands due to band - structure effects or a bilayer system with a small amount of hybridization ( the bonding and antibonding fermi surfaces are close ) .
we will assume that the system has full translation and rotation symmetries ( which will reduce to a point - group symmetry for a lattice system ) .
as before ( and for simplicity ) we will assume that the system is nonmagnetic ( so that the spin degrees of freedom yield only a redundant effect ) and that the fermi surfaces are such that the system is not unstable to the formation of charge density waves or any other instability .
we will also ignore possible superconducting states .
we will also assume that the separation between the two bands is small and only the forward - scattering channels need to be considered .
under these approximations , the scattering processes can be classified by their angular momentum channels , as in the case of the one - band model , and we can define bosonic fields ( fermion bilinears ) @xmath90 here @xmath91 stands for the normal - order product , relative to the ground state of a fermi liquid ( or gas ) . here
@xmath92 label the two bands .
the matrices @xmath93 are the identity matrix for @xmath94 and the pauli matrices for @xmath95 , @xmath96 , or @xmath97 .
@xmath98 is the polar angle of the momentum vector @xmath13 . with these definitions
, @xmath82 can be written as the sum of all quadratic terms in @xmath99 and @xmath100 that preserve momentum and angular momentum .
we further assume that the instability only occurs in one angular momentum channel @xmath83 so that collective excitations in all other angular momentum channels are gapped and irrelevant at low energies .
for now , we only consider one particular interaction hamiltonian @xmath82 of the form @xmath101 written in terms of fermionic operators , we can see that this interaction corresponds to the scattering channel @xmath102 .
other scattering channels will be studied later in sec .
[ sub : sec : other_scatterings ] .
in addition to the @xmath0 charge symmetry , this hamiltonian has an extra internal @xmath0 symmetry corresponding to the relative phase between the two bands .
this is because the interaction @xmath82 preserves particle number in each band .
this high symmetry requires some amount of fine tuning , but as we will show later , most of the properties are preserved even in the absence of this symmetry ( at least perturbatively ) .
just as in the case of the one - band model discussed in sec .
[ sec : single ] , if an interaction in some angular momentum channel is attractive and in magnitude exceeds a critical value , the ground state of the system becomes unstable .
the corresponding order parameters can be taken to be two two - component real vectors in the @xmath0 relative phase manifold @xmath103 with @xmath104 or @xmath105 .
notice that we use bold characters to represent vectors in space ( or momentum space ) but use @xmath106 to indicate the two two - component real vector order parameters which form a representation of the nondiagonal piece of @xmath107 group . in order to preserve the spatial symmetries and the internal @xmath0 symmetry
, the landau free energy has the form @xmath108 this free energy is very similar to the spin pomeranchuk instability states in ref .
@xcite , except that the internal symmetry here is the relative phase @xmath0 instead of the spin @xmath109 . the resulting mean - field phase diagram for the system at hand
is shown in fig .
[ fig : phase1 ] . the coefficients of the free energy can be determined by a mean - field calculation in the same spirit as that of ref.@xcite .
we obtain @xmath110 \\ u=&\frac{n(0)}{64 } \left[2\left(\frac{n^{\prime}(0)}{{n(0)}}\right)^2-\frac{n^{\prime\prime}(0)}{{n(0)}}\right ] , \label{eq : mean_field_u } \\
v=&\frac{n^{\prime\prime}(0)}{48}. \label{eq : mean_field_v}\end{aligned}\ ] ] here @xmath111 is the energy splitting between the two bands , which is assumed to be much smaller than the fermi energy @xmath112 , @xmath113 ; @xmath86 is the density of states at the fermi surface calculated using the average dispersion relation @xmath114/2 $ ] .
@xmath115 and @xmath116 are the first- and second - order derivatives of the density of states @xmath117 at the fermi surface .
higher order terms will be needed for stability reasons if @xmath118 or @xmath119 . for simplicity
, we only consider @xmath120 and assume that the higher order term is @xmath121 with @xmath122 . we will now discuss the structure of the phase diagram of fig.[fig : phase1 ] .
the system has three phases : a ) the normal phase with @xmath123 , b ) the @xmath124 phase in which @xmath125 , and c ) the @xmath15 phase in which they are orthogonal , @xmath126 and @xmath127 . for @xmath120 and @xmath128 , @xmath129 marks the second - order phase boundary between the normal state and either the @xmath124 or the @xmath15 phase , depending on the sign of @xmath130 . to distinguish this @xmath124 phase with other similar phases which will be discussed below
, we indicate this phase as @xmath131 in the phase diagram of fig .
[ fig : phase1 ] ( same with the @xmath15 phase ) .
this phase has a distorted fermi surface , as shown in fig .
[ fig : fs_alpha_1 ] . in the @xmath131 phase ,
the fermi surfaces of both bands are distorted , and their rotational symmetries are reduced from @xmath132 to a @xmath133-fold discrete symmetry .
the @xmath131 phase preserves the @xmath5 symmetry as well as @xmath10 and @xmath9 .
a similar phase on the square lattice is discussed in ref .
[ ] , where it is referred to as the `` hidden nematic phase . '' symmetry .
the red line is the phase boundary between the @xmath131 and the normal phases .
the green ones are between the @xmath134 and the normal phases , and the blue one is between the @xmath131 and the @xmath134 phases .
thick lines are first order phase boundaries and others are second - order ones .
the black dot is a tricritical point .
the dashed circle is a critical point where the first - order phase boundary meets two second - order phase transitions.,scaledwidth=40.0% ] for @xmath135 , the system is in the @xmath134 phase . in this phase ,
the fermi surfaces of the two bands remain isotropic with shifted fermi wave vectors , but the relative phases between the two bands are locked to each other . the relative phase changes by @xmath136 around the fermi surface , as shown in fig .
[ fig : fs_beta_1 ] . from a topological point of view , the @xmath134 phase is a map from the fermi surface @xmath137 to the @xmath137 manifold of the @xmath0 symmetry group .
the non - trivial homotopy group @xmath17 of this mapping is described by the kronecker index , the winding number , which is the angular momentum quantum number @xmath83 . under the @xmath5 or @xmath9 transformation , the kronecker index changes sign .
therefore , the @xmath134 phase breaks the @xmath5 and @xmath9 symmetries but preserves their combination .
hence , the @xmath134 phase is a type @xmath4 state and the topological nature of the kronecker index guarantees that the two degenerate @xmath5 symmetry - breaking ground states of the @xmath134 phase can not be transformed into each other by any continuous processes .
the @xmath138 line marks the first - order phase boundary between the @xmath131 and @xmath134 phases .
when @xmath119 , a first - order phase transition to the @xmath134 phase occurs with decreasing @xmath63 and this first - order phase boundary meets the second - order one at a tricritical point , the black dot in fig .
[ fig : phase1 ] .
the @xmath134 phase is a type @xmath4 @xmath5-breaking state with a nonzero spontaneous hall conductance @xmath139 . as shown in appendix [ app : hall_cond ] , following the results of haldane,@xcite the value of the hall conductance @xmath139 is quantized for an insulator but not for a conductor such as the @xmath134 phase .
however , we can still relate @xmath139 with the kronecker index of @xmath16 .
applying eq .
, for the @xmath134 phase , the integration region of the integral is the annular region comprised between the two fermi surfaces of the two bands .
if the energy difference between the two bands is small , the @xmath97 component of @xmath140 defined in eq . of appendix [ app : wilson_loop ] can be taken as a constant . under this approximation ,
the hall conductance is @xmath141 where @xmath142 and @xmath143 can be considered as the @xmath144 and @xmath96 components of a two - dimensional unit vector ( by definition , @xmath145 in @xmath15 phase ) .
the integral above is taken around the fermi surface and measures the kronecker index of @xmath16 , which counts the number of times the relative phase winds around the fermi surface .
notice that the prefactor of the integral , @xmath146 is unquantized and can be changed continuously .
hence , as expected , @xmath139 is not quantized in this phase . in particular , @xmath139 vanishes if the fermi surfaces coincide , @xmath147 .
we now study the effects of interactions that were not considered in sec . [
sub : sec : one_scattering ] .
the interactions that preserve the particle number in each band , for example , @xmath148 , can not change qualitatively any conclusions of sec . [
sub : sec : one_scattering ] , since these scattering processes preserve the relative @xmath0 symmetry of the two bands .
on the other hand , interactions that change the particle number of each band , such as @xmath149 or @xmath150 , break the relative @xmath0 symmetry and potentially can make a difference .
however , we can still distinguish the two different ordered phases , @xmath124 and @xmath15 , depending on whether @xmath151 vanishes or not . to distinguish from the @xmath131 and @xmath134 phases of sec .
[ sub : sec : one_scattering ] , we refer to the @xmath124 and @xmath15 phases for models with @xmath149 interactions but not @xmath150 interactions as the @xmath18 and @xmath19 states .
the corresponding phases when both these two kinds of processes are present will be denoted by @xmath152 and @xmath153 , respectively .
the @xmath152 and @xmath153 phases are the most general ones and do not require fine tuning .
the main issue we will be interested in is to determine if these are genuinely distinct phases , _
i.e. _ , if the order parameters have a different behavior in all of these cases
. similar to the @xmath131 phases , the @xmath18 and @xmath152 phases preserve the @xmath9 symmetry . as shown in figs .
[ fig : fs_alpha_2 ] and [ fig : fs_alpha_3 ] , the fermi surface of the @xmath18 phase has the same @xmath133-fold rotational symmetry as @xmath131 , but the @xmath152 phase has a lower , @xmath83-fold , rotational symmetry .
this is because the @xmath150 scattering processes couple the order parameters @xmath154 and @xmath155 to the order parameter of the charge pomeranchuk instability in each band @xmath156 where @xmath23 is the band index .
hence , @xmath157 and @xmath158 also acquire an expectation value in the @xmath152 phase , which reduces the rotational symmetries from @xmath133-fold down to @xmath83-fold .
therefore , for @xmath83 odd , the @xmath152 phase is a type @xmath3 time - reversal symmetry - breaking state .
the @xmath19 and @xmath153 phases break both @xmath5 and @xmath9 symmetries but preserve @xmath12 , and hence are type @xmath4 time - reversal symmetry - breaking phases .
this conclusion becomes obvious if one notices that the @xmath5 and @xmath9 symmetry breakings in the @xmath15 phases is described by a topological index as shown in eq . .
hence , this property should survive even after adiabatically turning on scattering processes that do not preserve the relative @xmath0 symmetry .
however , due to the absence of an exact relative @xmath0 symmetry , the symmetry between @xmath154 and @xmath155 is no longer preserved . as a result
, these two order parameters can not become critical at the same time as one tunes the control parameters .
hence , in order to reach the @xmath19 or @xmath153 phases from the normal fermi - liquid phase , the system must necessarily either go through a sequence of two phase transitions , at which one order parameter at a time will get a non - zero expectation value , or there will be a direct first - order transition to a state in which both are nonzero .
in addition , in the @xmath19 or @xmath153 phases , although the two order parameters @xmath159 and @xmath160 are still perpendicular to each other , without the protection of the @xmath0 relative phase symmetry their magnitudes are no longer equal .
hence , in these phases the fermi surfaces are no longer isotropic .
the @xmath19 phase with angular momentum channel @xmath83 has a fermi surface with @xmath133-fold rotation symmetry , as shown in fig .
[ fig : fs_beta_2 ] .
this phase breaks the @xmath5 and @xmath9 symmetries as the @xmath134 phase does but also has fermi surface with the same @xmath133-fold rotational symmetry as the @xmath131 and @xmath18 phase . in particularly , in the @xmath161 channel ,
the @xmath19 phase is a charge nematic state @xcite but with broken @xmath5 and @xmath9 symmetry . in the @xmath153 phase ,
much as in the case of the @xmath152 state , the @xmath162 scattering processes reduce the rotational symmetry to @xmath83-fold , as shown in fig . [
fig : fs_beta_3 ] .
hence , the @xmath87 @xmath153 phase is a charge nematic state with type @xmath4 @xmath5 symmetry breaking .
we studied the system without the @xmath0 relative phase symmetry by adding the following interactions @xmath163 \nn\\ & + \sum_{\mathbf{q};i=1,2}\frac{g_\ell^{(2)}(q)}{2}\left[\phi_{\ell i , x}(\mathbf{q})\phi_{\ell i , y}(\mathbf{-q})+ \phi_{\ell i , y}(\mathbf{q})\phi_{\ell i , x}(\mathbf{-q})\right].\end{aligned}\ ] ] which correspond to the @xmath164 scattering processes . in mean - field theory , the landau free energy becomes @xmath165 with @xmath166 and @xmath130 the same as in eqs . and and @xmath167 \nn\\ & -\left(\frac{1}{4 \left(f_\ell(0)+|g_\ell(0)|\right)}+\frac{1}{4 \left(f_\ell(0)-|g_\ell(0)|\right)}\right ) \\
\delta=&-\frac{1}{4 \left(f_\ell(0)+|g_\ell(0)|\right)}+\frac{1}{4 \left(f_\ell(0)-|g_\ell(0)|\right)}. \label{eq : anisotropic - parameters}\end{aligned}\ ] ] here @xmath168 .
in fact , a complex @xmath169 adds terms such as @xmath170 to the landau free energy , but upon a suitable rotation , the landau free energy can be transformed into the form of eq .. the notation is the same as before and the leading higher order term is assumed to be @xmath171 with @xmath122 for simplicity . symmetry .
the notation is the same as in fig.[fig : phase1 ] .
the circle represents a bicritical point where a second - order phase boundary and two first order boundaries meet.,scaledwidth=40.0% ] the phase diagram is shown in fig .
[ fig : phase2 ] . as predicted by the general symmetry arguments
, the @xmath18 phase can be reached through a second - order phase transition at @xmath172 . to reach the @xmath19 phase one must either go through two transitions ( using the @xmath18 phase as an intermediate phase ) or by a direct first - order transition .
the transition between the @xmath18 and @xmath19 phases may be first order or second order depending on details . in this mean - field theory
the first - order and second - order phase boundaries meet at a tricritical point ( the black dot in fig .
[ fig : phase2 ] ) located at @xmath173 and @xmath174 .
the first - order phase boundary between @xmath18 and @xmath19 phases meet the second - order phase boundary between @xmath18 and the normal phases , as well as the first - order phase boundary between @xmath19 and the normal phases , at a bicritical point ( the circle in fig .
[ fig : phase2 ] ) . for most general interactions
, the @xmath152 and @xmath153 phases can also occur .
the @xmath152 and @xmath153 phases have a similar phase diagram as the @xmath18 and @xmath19 ( fig .
[ fig : phase2 ] ) and do not exhibit any essentially new phenomena .
hence , we will not present here the mean - field study for the @xmath152 and @xmath153 phases , which has a similar structure to what we have already described in this section .
most of the conclusions we reached in the continuum models of the previous section can be generalized to the case of lattice models in the absence of degeneracy point .
one principal difference in the case of lattice models is that the continuum rotational symmetry is broken down to a discrete point - group symmetry .
in particular this gaps out the corresponding goldstone modes .
another one is that in lattice models the quantum phase transitions to nematic states ( and their generalizations ) often ( although this is not necessarily always the case as there are a few known counterexamples ) also involve a topological lifshitz transition ( from closed to open fermi surfaces ) leading to first - order quantum phase transitions.@xcite for one - band models , obviously , the band structure has no degeneracy points .
same as in the continuous model , only the type @xmath3 @xmath5 broken - symmetry state can exist . as an example , we study a one - band model on a simple triangular lattice . this lattice has a @xmath175-fold rotational symmetry . with properly chosen interactions ,
the system can undergo a pomeranchuk instability and form the flux state shown in fig . [
fig : triangular ] . in this flux state ,
three currents flow along the three bonds of the simple triangular lattice . as a result
, there is a positive flux in each up - pointing triangle and a negative flux in each down - pointing triangle .
the net current and the total flux in each unit cell are both zero , but the rotational symmetry of the fermi surface will in general be reduced from @xmath175-fold to @xmath176-fold symmetry , except when the flux in each triangle is @xmath59 for integer @xmath23 .
this is so because a gauge transformation can change the flux in a triangle by @xmath177 and similar effects is known to occur for the square lattice .
@xcite the same analysis with very little modification applies to the `` varma loop model '' @xmath70 ( ref .
@xcite ) and to a model on the square lattice with effective diagonal hopping terms and the same pattern of time - reversal symmetry breaking recently discussed in ref . which are also type @xmath3 time - reversal symmetry - breaking states . when the flux is not @xmath59 , @xmath178 , where @xmath179 is the identity operator . as shown in fig .
[ fig : triangular ] , the fermi surface of a state with flux different from @xmath59 has @xmath176-fold rotational symmetry , which corresponds to the @xmath69 pomeranchuk instability , and this state belongs to the type @xmath3 @xmath5 symmetry - breaking states . in two - band lattice models without degeneracy points , @xmath124 and @xmath15 phases can also occur much as in the continuum model discussed before .
the simplest example is a bilayer model in which two layers of a lattice system are separated by a small distance .
if the separation is small enough , three dimensional ( 3d ) effects can be ignored . for a bilayer model , two bands ( bonding and antibonding )
can be formed and they usually have no degeneracy . in ref .
@xcite , puetter , doh , and kee studied a system of this type and found a `` hidden nematic phase '' on a bilayer square lattice , which is one of the example of the @xmath131 phase . for band structures with degeneracy points ,
the type @xmath4 phase just requires one order parameter quadratic in fermions .
a well - known example is the flux states in the honeycomb lattice .
the band structure of the honeycomb lattice has two degeneracy points ( the dirac nodal points ) at the two corners of its first brillouin zone .
as shown in appendix [ app : wilson_loop ] and fig .
[ fig : flux_honeycomb ] , there is a `` monopole flux '' @xmath180 passing through each dirac point .
if the two dirac points get masses with the same sign , they can not cancel each other ; the @xmath9 and @xmath5 symmetries are broken , as shown in fig . 1 of ref .
( a similar effect is found in the mean - field theory of the chiral spin liquid @xcite . )
a recent mean - field study of a system of interacting fermions on the honeycomb lattice ( at half filling ) shows that this state can be stabilized by repulsive next - nearest - neighbor interactions .
another example is a fermionic system on the crossed - chain lattice as shown in fig .
[ fig : crossed_chain ] .
this lattice has two sublattices.@xcite the tight - binding model on this lattice has two bands and one degeneracy point in each brillouin zone where the two bands touch .
this degeneracy point has monopole flux @xmath181 [ fig .
[ fig : flux_crossed_chain ] ] . if the four - fold rotation symmetry of the lattice is broken explicitly upon the introduction of a different chemical potential for each sublattice , or spontaneously via a ( quantum nematic ) symmetry breaking , the degeneracy point with flux @xmath181 splits into two dirac points , each with flux @xmath180 .
similar as in the case of the honeycomb lattice , a dirac mass term which removes the band crossing breaks the @xmath5 and @xmath9 symmetries .
this state corresponds to the flux state shown in fig .
[ fig : crossed_chain ] with order parameter @xmath182 where @xmath183 , @xmath184 , @xmath185 , and @xmath186 are the creation and annihilation operators for the two sublattices of the crossed - chain lattice at momentum @xmath187 .
it can be shown that this is a type @xmath4 time - reversal symmetry - breaking state which can be stabilized by an arbitrarily weak nearest - neighbor repulsive interaction for some range of hopping amplitudes and electron densities.@xcite recently , ran and coworkers @xcite , found similar degeneracy points with @xmath188 monopole flux in feas - based materials with @xmath5 symmetry . hence
, similar @xmath5 symmetry - breaking phases may be possible in such systems as well . the state shown in fig .
[ fig : crossed_chain ] has a close analogy to the @xmath78dw state with @xmath189 symmetry [ fig .
2.(c ) in ref . @xcite and fig . 1 in ref .
both phases are type ii @xmath5 symmetry - breaking states .
they have similar order parameters but with very different physical origins .
first of all , the lattice structure of our model is a crossed - chain lattice while the @xmath78dw state is defined on a simple square lattice .
hence , the alternating of the diagonal hopping strength ( the crosses ) in our model is due to a lattice effect ( an explicit symmetry breaking ) , while in @xmath78dw state it is due to spontaneous translational - symmetry breaking .
secondly , due to the special lattice structure of the crossed - chain lattice , a non - generic ( non - dirac ) band crossing is presented at momentum @xmath190 in the absence of interactions .
this band touching leads to an infinitesimal instability to the @xmath5 symmetry - breaking flux state shown in fig .
[ fig : crossed_chain ] , where the @xmath189 @xmath78dw state requires a finite interaction to be reached .
we end the discussion on lattice models with a few remarks on the emery model , which describes the @xmath191 plane of cuprates .
the varma loop @xmath70 state , shown in fig .
[ fig : varma_loop ] , is a type @xmath3 state that breaks the @xmath10 and @xmath5 but not the @xmath9 and @xmath11.@xcite without breaking translational and charge @xmath0 symmetries , we show in figs .
[ fig : emery_1 ] and [ fig : emery_2 ] two other nonmagnetic states that break the @xmath9 symmetry as well as @xmath5 .
the state shown in fig .
[ fig : emery_1 ] is the same as the state in fig .
[ fig : crossed_chain ] , if we notice that the oxygens ( @xmath192 and @xmath193 orbitals ) in the emery model form a crossed - chain lattice ( rotated by @xmath194 ) .
the state in fig .
[ fig : emery_2 ] involves three bands and there is a flux piercing each small triangle formed by neighboring @xmath195 , @xmath192 , and @xmath193 orbitals . in each unit cell , three of the triangles have flux @xmath196 but the fourth triangle has flux @xmath197 , with zero total flux in the unit cell . in this state
all the @xmath5 , @xmath9 and @xmath10 symmetries are broken ( and none of the pairs @xmath75 , @xmath71 , or @xmath198 is preserved ) while @xmath199 remains unbroken . hence
, this state is a mixture of states of type @xmath3 and @xmath4 time - reversal symmetry breaking .
for a two - band model for a system in the continuum ( _ i.e. _ , ignoring the explicit breaking of rotational symmetry by the underlying lattice ) , the @xmath131 and @xmath134 phases break spontaneously two continuous symmetries : the @xmath132 rotational invariance and the relative @xmath0 phase symmetry .
hence , two goldstone modes are generated in the broken - symmetry phase . at finite temperatures ,
the thermal fluctuations of these two goldstone modes destroy this long - range order .
the thermodynamic phase transition is in the universality class of a system of two @xmath200 models which will undergo a kosterlitz - thouless type phase transition . for the case of the @xmath18 and @xmath152 phases , the relative @xmath0 phase symmetry of the bands
is broken explicitly by the effects of the interactions , which do not preserve particle number on each band as a good quantum number .
hence , only one continuous symmetry , @xmath132 , remains and it is broken spontaneously in this phase at @xmath201 . since this is a continuous @xmath132 symmetry , the finite temperature transition between the normal state and the @xmath18 or the @xmath152 phase also belongs to the kosterlitz - thouless ( kt ) universality class , in this case with a single @xmath132 order - parameter field . the direct transition from the normal state to the @xmath19 or @xmath153 states was shown in sec .
[ sub : sec : other_scatterings ] to be first order at zero temperature .
therefore one expects this transition to remain first order even at finite temperature up to a critical value where the phase boundary reaches a tricritical ( or multicritical ) point . in general , for a system with full rotational invariance @xmath132 , the transition from these phases to the normal state should be in the kt universality class as well . for lattice systems ,
rotational symmetry is broken down to a discrete subgroup , the point - group symmetry of the underlying lattice .
hence the only continuous symmetry left to be spontaneously broken in the @xmath131 and @xmath134 phases is the relative @xmath0 phase .
thus , the finite temperature transitions from the @xmath131 and @xmath134 phases to the normal state are also kt transitions . for lattice systems
in which the @xmath18 , @xmath19 and @xmath152 , @xmath153 phases can be realized there are no continuous symmetries present since the relative @xmath0 symmetry is broken by the interactions . hence ,
in general the thermodynamic transitions from the @xmath19 and @xmath153 phases to the normal state only involve the restoration of time - reversal invariance and the discrete point - group symmetries broken in these low - temperature phases ( up to important caveats discussed below ) . for a lattice system , except for the @xmath131 and @xmath134 phases ,
whose existence is not generic and requires fine tuning , the @xmath124 and @xmath15 phases only break discrete symmetries and hence do not have goldstone modes in their excitation spectra . as a result ,
the fermionic quasiparticles are the only low - energy excitations in these broken - symmetry phases .
this implies that the fermi - liquid picture should remain valid in these low - temperature phases even if the fluctuations are stronger than what is allowed in a naive mean - field treatment
. we will now give a more careful analysis of the nature of the thermal phase transitions for the @xmath18 , @xmath19 , @xmath152 , and @xmath153 phases in lattice , which have richer structures , based on an analysis of the symmetry .
let us begin by discussing the d - wave type @xmath153 phase in a square lattice . in this case , in addition to the @xmath74 time - reversal symmetry
, this phase also lowers the invariance under @xmath202 spatial rotation to @xmath203 rotation ( 2d space inversion ) , which is also a @xmath74 symmetry breaking .
formally , the @xmath153 phase has a broken @xmath204 symmetry . in the case of the @xmath19 phase ,
the situation is similar except that the time - reversal even @xmath74 symmetry requires a different physical interpretation ( see appendix [ app : z2 ] ) . from this view point the natural description of the thermal phase transitions out of the @xmath19 or @xmath153 phase should be in the 2d ashkin - teller universality class . in what follows we will consider the conceptually more straightforward @xmath153 phase , but the same analysis also applies to the more intricate @xmath19 phase .
if the low - temperature phase is @xmath153 the phase diagram can have the general topology of the one shown in fig .
[ fig : phase - thermal1 ] . as temperature
is raised it may undergo a direct phase transition to the normal phase or have two thermal transitions . in the latter case ,
the intermediate phase arises from the restoration of one of the @xmath74 symmetries , either time reversal or spatial rotations .
thus the intermediate temperature phase is either a time - reversal even nematic phase ( the @xmath152 phase ) or a spatially isotropic phase with broken time - reversal invariance . in any case , the transition between these two states in two space dimensions is continuous and in the 2d ising universality class .
one should note that it is possible to conceive the existence of a spatially isotropic intermediate phase with broken time - reversal symmetry even at @xmath201 . however
a state like that can not be reached within mean - field theory as it results from the `` quantum melting '' of one of the @xmath74 states of the @xmath15 phase .
these arguments suggest that the finite temperature critical behavior of this system is in general describable by a 2d classical ashkin - teller model , as far as the thermal transitions are concerned .
indeed this would be the case if the two ising transitions were to meet at a multicritical point , which would necessarily be in the 2d four - state potts model universality class , a kt transition .
if this scenario is correct , the transition from the @xmath15 phase to the normal state should have a line of continuous transitions with varying exponents .
since the quantum phase transition ( at @xmath201 ) between the normal state and the @xmath15 phase is first order , this scenario requires the existence of a tricritical point at some intermediate but low temperature .
a similar phase diagram was found in interacting monomer - dimer models on a 2d square lattice.@xcite an alternative possible scenario is that the two ising transitions do not meet at a multicritical point . instead , that the lower temperature ising transition becomes first order at some value of the control parameter , _
i.e. _ , a 2d ising tricritical point .
the general topology of the phase diagram is depicted in fig.[fig : phase - thermal2 ] . here
too the intermediate phase can either be an @xmath124 phase , in which case time reversal is restored at the lower temperature transition and the intermediate phase is nematic , or isotropy is restored first and the intermediate phase breaks time - reversal . still a third possibility arises if at @xmath201 the @xmath15 and normal phases are separated by a region of the @xmath124 phase . in this case
the @xmath201 transition is continuous , and one would generally expect two thermal transitions , as shown in fig.[fig : phase - thermal3 ] , with an intermediate temperature phase that at least at low temperatures must be an @xmath124 phase . the possibility of another intermediate phase discussed above , isotropic and @xmath5 breaking , can not be excluded even in this scenario .
finally , let us consider briefly the case of @xmath205 lattice symmetry breaking , beyond the @xmath206 symmetry of the square lattice we have been discussing here , _
i.e. _ , simple triangular and honeycomb lattices .
we will now have to consider more general angular momentum channels , with @xmath83 even or odd . specifically , for the simple triangular and honeycomb lattices the simplest cases of interest have @xmath87 and @xmath69 broken by the @xmath205 symmetry .
an example of this is the one - band model on a simple triangular lattice discussed above .
two - band models on these lattices allow for a richer structure .
the simplest case is a @xmath153 state with an @xmath69 particle - hole condensate without relative particle - number conservation of any type .
this state breaks spontaneously the @xmath74 inversion symmetry and the @xmath74 chiral symmetry .
hence , the thermal transitions to the normal state of this phase also have a @xmath204 , as in the square lattice case discussed above , with a phase diagram similar to those in figs.[fig : phase - thermal1]-[fig : phase - thermal3 ] .
( just as in the case of the square lattice , this analysis also applies to the phase @xmath19 with the same caveats on the interpretation of the symmetries . )
a different symmetry - breaking pattern arises on both the simple triangular and honeycomb lattices if the condensate is in the @xmath87 particle - hole channel . in the @xmath153 state with @xmath87
, the @xmath207 symmetry is lowered to @xmath208 ( inversion in 2d ) symmetry , where the @xmath209 axis is lifted .
hence , @xmath210 is the broken rotational symmetry manifold of the order parameter .
in addition to this @xmath211 , this phase also breaks the @xmath74 time - reversal symmetry , which here is equivalent to chiral symmetry .
hence , it breaks a @xmath212 symmetry .
the thermal phase diagram for this @xmath213 problem is more complex ( and old problem ) than the cases we presented above ( see refs .
[ ] and [ ] and references therein ) .
the structure of the phase diagram can be summarized in three cases : a ) a direct transition to the normal state , b ) an intermediate temperature critical phase , and c ) an intermediate phase with long - range order . in the first case ,
the transition between the @xmath153 phase and the normal phase is a direct first - order transition ( similar to that of the closely related six - state potts model ) . in the second case
, there is a finite range of temperatures in which the system is critical , as in the @xmath214 model , and has two kt - type transitions at each end - point .
the third case consists of a sequence of partial restorations of the @xmath211 and @xmath74 broken symmetries of the @xmath153 state through intermediate temperature phases with either @xmath211 nematic order and no broken time - reversal symmetry or with broken time - reversal invariance and full isotropy .
we studied microscopic fermionic models ( generally in metallic phases ) with spontaneous breaking of time - reversal invariance .
we considered one - band models and two - band models ( ignoring spin ) with and without separate conservation of particle number in each band . the time - reversal breaking phases can be classified in two classes : 1 .
type @xmath3 phases , which break time reversal ( @xmath5 ) and space inversion ( @xmath10 ) but do not break chirality ( @xmath9 ) .
we found that type @xmath3 phases occur in both one - band and multiple - band fermionic systems in a generalized nematic ground state with a ( particle - hole ) condensate in an odd angular momentum channel .
examples of type @xmath3 phases we discuss are one - band nematics with @xmath69 and two - band models in the @xmath152 state , which has a particle - hole condensate with @xmath83 odd without independent particle number conservation in each band .
type @xmath4 phases , which break time reversal ( @xmath5 ) and chirality ( @xmath9 ) but do not break space inversion ( @xmath10 ) .
type @xmath4 phases have a spontaneous ( non - quantized ) anomalous hall effect .
we found that type @xmath4 phases are not realized in one - band models . in two ( and multiple ) band models they occur in interband particle - hole condensates with _ even _ angular momentum if particle number is not separately conserved in each band ( @xmath153 phases ) , @xcite and in any angular momentum channel @xmath215 provided particle number is either preserved in each band ( @xmath134 phases ) or conserved modulo @xmath105 ( @xmath19 phases ) .
each class of states has unique experimental signatures that can be detected in linear and nonlinear conductivity measurements , and in the optical response with polarized light .
let us discuss first how these phases can be detected in transport .
it is well known that electronic nematic phases induce an anisotropy in the conductivity tensor.@xcite in a metallic system , an in - plane electric field @xmath179 induces a current @xmath216 which can be expanded as a power series in the electric field @xmath179 , @xmath217 with @xmath218 , @xmath219 , and @xmath220 being @xmath144 or @xmath96 . the first term in eq .
is the linear response , and the second term is the leading nonlinear response .
the different components of the conductivity tensor ( and of the nonlinear response ) can be arranged to transform properly under chiral @xmath9 and space inversion @xmath10 symmetries , and can be used to detect these broken - symmetry phases in experiment .
thus , the conductivity tensor is sensitive to both rotations under @xmath221 and chirality , while the third rank nonlinear conductivity tensor is odd under inversion . to detect systems with condensates with @xmath222 ,
it is necessary to consider higher order nonlinear response terms .
thus , one can construct phenomenological `` order parameters '' using the electrical response tensors , and the simplest ones are presented in table [ table : conductance ] . the signatures of type @xmath3 @xmath223 breaking phases can be detected through nonlinear optical processes , such as polarized raman scattering .
[ they can also be checked directly by angle - resolved photoemission spectroscopy ( arpes ) by detecting the anisotropy of the fermi surface .
] on the other hand , type @xmath4 phases , such as the @xmath15 phases , can be detected optically through a nonzero kerr effect ( in the absence of external magnetic fields ) .
.conductance tensors for different phases . [ cols="^,^,^,^,^,^,^,^",options="header " , ] the problem of constructing phases of electronic systems with broken time - reversal invariance was presented here mainly from its intrinsic conceptual interest .
our interest in this problem has been to a large extent motivated by the recent discovery of time - reversal symmetry - breaking effects in the ruthenates and in the cuprates .
we should stress that although the theory we presented in this paper as it stand can not describe a strongly correlated system , the patterns of symmetry breaking , as well as their consequences , should be of more general validity .
the strongest experimental evidence available to date of time - reversal symmetry breaking in superconductors is in the layered compound .
kerr effect rotation experiments@xcite and corner junction experiments@xcite strongly suggest that this material may indeed have a @xmath224-wave superconducting state which breaks spontaneously time - reversal symmetry .
this is consistent with a theoretical prediction@xcite of a condensate with a @xmath225 order - parameter symmetry .
this evidence is however not fully uncontroversial given the conflicting experimental results of ref . [ ]
which , so far , have failed to detect the expected edge currents of the @xmath76 superconductor .
recent high - precision kerr rotation experiments@xcite have now given evidence of weak but detectable time - reversal symmetry breaking in the underdoped pseudogap regime of the high @xmath6 compound .
neutron - scattering experiments in underdoped @xcite , and in hgba@xmath226cuo@xmath227 @xcite similarly suggest that the breaking of time - reversal invariance may also occur in these materials .
these recent discoveries in the cuprates and in have renewed interest in the possible mechanisms of time - reversal symmetry breaking in strongly correlated systems .
however , it is worth to note that , in addition to the noted evidence of time - reversal symmetry breaking , neutron - scattering experiments@xcite and transport experiments@xcite find strong evidence for nematic charge order in in the same doping range .
it is unreasonable to believe that these two phases can be unrelated to each other , and perhaps they have a common origin . in this sense
, the phases found in this paper may shed some light on these issues .
one important problem that we have not discussed in this work is the role of disorder in these phases .
it is well known that disorder couples as a _ random field _ to the order parameter of anisotropic nematic - like phases,@xcite destroying the ordered state and rendering the system glassy .
the same applies to other phases we discussed here that break spontaneously either point - group symmetries and/or inversion . on its own
, nonmagnetic disorder can not couple directly to the chiral symmetry and does not destroy automatically a type @xmath4 time - reversal breaking state .
however , if the spin degrees of freedom are also included even nonmagnetic disorder can couple indirectly to time - reversal breaking order parameters through the effects of spin - orbit interactions . in this case
the system becomes a @xmath5-breaking glassy state . on the other hand
, disorder can induce @xmath5 and @xmath9 breaking effects in phases such as type @xmath3 states by breaking locally translation , inversion , and point - group symmetries of the system . in any case
, the time - reversal symmetry - breaking effects induced by disorder should be quite weak.@xcite other states that break time - reversal symmetry to varying degrees have been postulated in the context of high @xmath6 compounds .
these states assume the existence of spontaneously circulating currents in the ground state .
they include the loop state advocated by varma , @xcite which breaks time reversal but not translation invariance , and the @xmath78-density wave state of charkravarty , laughlin , morr and nayak @xcite , a @xmath78dw state that breaks time reversal and translation invariance ( by one lattice spacing ) but it is invariant under the simultaneous action of both transformations . in the absence of disorder neither of these states exhibits a uniform kerr effect .
et al_@xcite proposed a @xmath78dw state with @xmath189 symmetry that breaks translational symmetry and has nonzero kerr effect .
there is a close analogy between the @xmath15 phase and the @xmath225 ( or @xmath77 ) superconducting state .
they both need two real order parameters that couples the @xmath132 rotational symmetry and an internal @xmath0 symmetry to break the @xmath5 and @xmath9 symmetries .
the @xmath0 symmetry in the @xmath225 spin triplet condensate in the particle - particle channel in superconductors is the charge ( `` gauge '' ) @xmath0 symmetry , which is an exact symmetry of the system and can not be broken explicitly .
on the other hand , the formally analogous @xmath15 phase is a particle - hole condensate which breaks spontaneously the relative @xmath0 phase symmetry of a multiband system .
this symmetry in general is not exact but it is asymptotically exact `` emergent '' symmetry at the fermi - liquid fixed point , since the symmetry - breaking terms are formally irrelevant ( `` dangerous irrelevant '' ) operators which are always present in any real system .
the difference , between exact and `` emergent '' symmetries , has a direct consequence on the structure of the phase diagram , rendering the quantum phase transition from the normal state to the @xmath15 phase first order or through an intermediate @xmath124 phase .
our @xmath15 phases have a close similarity with the phases ( with the same name ! ) in spin-1/2 fermionic systems with anisotropic phases such as those discussed in refs . [ ] and [ ] .
however , the @xmath15 phases of those systems are not type @xmath4 @xmath5 symmetry - breaking states for two reasons .
first under time reversal all three components of the spin polarization of the quasiparticles changes sign .
in contrast , here we have used the pauli matrices to act on an internal space unrelated with the electron spin , a `` pseudospin '' representing the multiple electronic bands . in our case
_ only _ the complex , @xmath228 , component is odd under time reversal .
secondly , the internal symmetry in our problem is only an approximate @xmath0 phase symmetry , while in the spin problem it is the full @xmath109 group ( in the absence of the spin - orbit couplings ) .
we have shown that for systems with @xmath2 degenerate bands , the ( natural ) symmetry is @xmath229 ( @xmath230 for this case ) , and the berry connection @xmath231 , defined in sec . [
sec : gauge ] , is always trivial and it can always be eliminated by a gauge transformation ( it is a gauge transformation ! ) due to eq . .
hence , the actual berry phase [ no matter @xmath0 or @xmath109 ] is zero if the @xmath229 symmetry is exact . as a result , in the fully symmetric case ( as well as in the spin-1/2 model )
there are no type @xmath4 @xmath5 symmetry - breaking states .
this can also be seen if we notice that the map from the fermi surface to the internal @xmath0 group is @xmath16 , which has a nontrivial homotopy group , @xmath17 , with a non - vanishing kronecker index .
in contrast , for the case of spin-1/2 systems , the mapping @xmath233 has a trivial homotopy , @xmath234 , and does not have a nontrivial topological index .
we should emphasize that the electronic quasiparticles of the systems we have considered do carry spin-1/2 , but these degrees of freedom play no role in the phases we have discussed as they are paramagnetic .
we thank steven kivelson and congjun wu for many insightful discussions .
this work was supported in part by the national science foundation grant dmr 0758462 at the university of illinois ( ef ) and by the department of energy , division of basic energy sciences under award de - fg02 - 07er46453 through the frederick seitz materials research laboratory at uiuc ( ks and ef ) .
by definition , the matrix @xmath36 satisfies @xmath235 this constraint on @xmath36 implies that the off - diagonal terms and the diagonal terms are related .
it follows that @xmath236 which leads to the conclusion that there is no berry phase in the overall ( charge ) @xmath0 sector . for a two - band model
, we can construct three gauge invariant objects that are sensitive to time - reversal symmetry breaking @xmath237 however , due to eq .
, only one of them is linearly independent .
hence , in the study of the @xmath5 symmetry breaking in any two - band model , only the diagonal term @xmath238 ( or @xmath239 ) is needed to be considered .
the same conclusion is trivially valid for one - band models .
from a topological point of view , in a two - band model , a unitary transformation at momentum @xmath13 belongs to the group @xmath73 . by removing the @xmath89 gauge degrees of freedom ,
the physical degrees of freedom are in the manifold @xmath240 .
hence , for each closed loop @xmath241 in momentum space , a map @xmath242 can be defined .
similar to the coherent - state path integral of spin @xmath243 ( see for instance ref .
@xcite ) , this map leads to a wess - zumino term .
we will show here the berry phase studied in the main text is half of the wess - zumino term of this map ( mod @xmath244 ) .
for a two - band model , the kinetic energy part of the hamiltonian is a @xmath245 hermitian matrix , which can be expanded in the basis of the identity matrix and the three pauli matrices : @xmath246 away from degeneracy points , @xmath247 is nonzero .
hence , we can define a 3d unit vector @xmath248 from a topological point of view , this 3d unit vector field is a map from the momentum space ( with degeneracy points removed ) to @xmath249 .
if we define the polar coordinates of the unit vector @xmath140 as @xmath250 and @xmath251 , where @xmath252 , the hamiltonian can be diagonalized by a unitary transformation , @xmath253 , where @xmath254 here @xmath255 and @xmath256 is any function of the momentum , which reflects the @xmath89 gauge freedom in a two - band model . therefore , we get @xmath257 the loop integral around contour @xmath241 is @xmath258 where @xmath23 is an integer measuring the winding number of the angle @xmath259 . by comparison ,
the wess - zumino term of the map from @xmath241 to @xmath249 is @xmath260 here @xmath261 is an arbitrary 2d manifold whose boundary @xmath262 is @xmath241 and @xmath23 is an integer determined by the choice of @xmath261 .
hence , we conclude that the berry phase @xmath263 is half of the wess - zumino term up to @xmath177 . under a change in chirality , the wess - zumino term changes sign ( mod @xmath264 ) . therefore , @xmath265 changes into @xmath266 under @xmath9 , which is the same as @xmath11 .
hence , @xmath265 is invariant under @xmath8 .
for the hamiltonian shown in eq . , the hall conductance @xmath139 can be evaluated as in ref .
@xcite , using the three - component unit vectors @xmath267 defined in eq .
, @xmath268 here the integration region @xmath261 is the area in momentum space where one band is filled and the other is empty . for insulators , the region @xmath261 is the whole momentum space ( the first brillouin zone for lattice models ) , which is a compact manifold .
hence , @xmath139 is the kronecker index of the mapping from a compact manifold to @xmath249 , the first chern number , which is quantized to be an integer .
it measures how many times the compact manifold wraps around the sphere @xmath249 .
for conductors , the boundary of @xmath261 is the fermi surface , which implies that @xmath261 is not a compact manifold .
hence , for conductors in general @xmath139 is not quantized.@xcite using the conclusion from appendix [ app : wilson_loop ] , the hall conductance @xmath139 can be related to the loop integral @xmath263 as @xmath269 where the contour @xmath241 is the boundary of @xmath261 , which is the fermi surface , and we choose the gauge where @xmath238 is analytic in region @xmath261 to remove the @xmath177 uncertainty in @xmath263 .
this result is straightforward if we notice that @xmath263 contains all the information of @xmath5 and @xmath9 symmetry breaking in two - band models .
therefore , @xmath139 , which measures the type @xmath4 @xmath5 symmetry breaking at low energies , must be directly related with @xmath263 , where @xmath241 is the fermi surface , which dominates the low - energy physics .
for a @xmath78-wave - like ( @xmath87 ) @xmath18 phase on a square lattice , the symmetry breaking is reduced from @xmath270 to @xmath271 . in the normal phase , in addition to the point - group symmetry of the square lattice @xmath206
, there is a internal @xmath74 symmetry corresponding to the relative phase shift by @xmath203 between the two bands of a system , in which the relative particle number is conserved mod @xmath105 , and the relative @xmath0 phase symmetry is thus reduced to phase shifts of the electron wave functions between the two bands by @xmath203 , @xmath272 and @xmath273 . in the @xmath18 phase , the order parameter is blind under the simultaneous action of a @xmath202 spatial rotation and the relative phase shift by @xmath203 .
hence , the resulting symmetry of the ordered phase is @xmath274 .
therefore , the thermal phase transition from the @xmath18 state to the normal state is in the 2d ising universality class .
the same analysis applies to the @xmath152 phase , except that the broken @xmath74 symmetry are simply @xmath275 , which only involves the spatial symmetry . for the @xmath19 phase , in addition to the @xmath74 symmetry breaking
just discussed above , an additional @xmath74 time - reversal symmetry is broken .
hence , the @xmath19 phase has a broken @xmath276 symmetry . | we study the spontaneous non - magnetic time - reversal symmetry breaking in a two - dimensional fermi liquid without breaking either the translation symmetry or the @xmath0 charge symmetry . assuming that the low - energy physics is described by fermionic quasiparticle excitations
, we identified an `` emergent '' local @xmath1 symmetry in momentum space for an @xmath2-band model . for a large class of models , including all one - band and two - band models
, we found that the time - reversal and chiral symmetry breaking can be described by the @xmath1 gauge theory associated with this emergent local @xmath1 symmetry .
this conclusion enables the classification of the time - reversal symmetry - breaking states as types @xmath3 and @xmath4 , depending on the type of accompanying spatial symmetry breaking .
the properties of each class are studied . in particular , we show that the states breaking both time - reversal and chiral symmetries are described by spontaneously generated berry phases .
we also show examples of the time - reversal symmetry - breaking phases in several different microscopically motivated models and calculate their associated hall conductance within a mean - field approximation .
the fermionic nematic phase with time - reversal symmetry breaking is also presented and the possible realizations in strongly correlated models such as the emery model are discussed . |
the protein constituents of the cytoskeleton of eukaryotic cells can be divided broadly into the following three categories : ( i)_filamentous _ proteins , ( ii ) _ motor _ proteins , and ( iii)_accessory _ proteins .
the three classes of filamentous proteins which form the main scaffolding of the cytoskeleton , are : ( a)_actin _ , ( b)_microtubule _ , and ( c)_intermediate filaments_. the three superfamilies of motor proteins are : ( i)_myosin _ superfamily , ( ii)_kinesin _ superfamily , and ( iii)_dynein _ superfamily .
microtubules serve as tracks for kinesins and dyneins.@xcite .
the main cellular mission of the molecular motor kinesin-5 is to cross - link anti - parallel microtubules and to slide them apart , thus playing a critical role during bipolar spindle formation @xcite .
kinesin-5 is absolutely essential for the dynamic assembly and function of the mitotic and meiotic spindles @xcite . in this work ,
i have closely followed the paper by peskin and oster @xcite .
the molecular motor kinesin-5 is a four headed motor with two heads on each side .
fig-[schematic ] shows a schematic representation of the model we have developed for this motor . in our model
, we assume the two heads of the motor on one side to be joined at a hinge .
similarly , the other two heads on the other side are also joined by another hinge .
these two hinges are then linked by a flexible connector .
we also assume that the two heads on each side are free to rotate about their respective hinge .
the two heads can also pass each other freely .
however there is one restriction on their relative displacement .
the maximum separation between two heads on one side can be equal to the separation between two discrete binding sites on the microtubule .
this separation , which is a constant , is denoted by @xmath0 .
we assume that the microtubules are aligned anti - parallel because kinesin-5 has a much stronger preferences for cross - linking anti - parallel overlapped microtubules than parallel ones @xcite .
the @xmath1 end directed stepping activity of kinesin-5 @xcite slides microtubules away from each other .
fig.[pair ] shows schematically the model of one pair of heads on a microtubule .
the factors included in the model which are responsible for the motor to move forward are : ( i ) the back head has a higher probability to detach than the front head .
( ii ) the attached head prefers a geometry in which it leans in the forward direction i.e towards the @xmath1 end of the microtubule .
the first of these factors ensures that the back head detaches and the second one ensures that it reattaches as a front head @xcite .
when both the heads are attached to the microtubules then the back head is in a more favourable orientation than the front one .
therefore , atp hydrolysis and subsequent detachment is more likely at the back head .
once the back head detaches , the front head relieves itself of the strain by leaning forward and swinging the detached back head forward .
hence this is the process by which the motor moves forward .
the coordinates of the two heads are taken to be @xmath2 and @xmath3 .
the hinge is located halfway betwween them .
the coordinate of the hinge , @xmath4 is therefore @xmath5 and the restriction dictates @xmath6 let [ cols="<,^ , < " , ] @xmath7 = probability that the free head binds behind the bound head .
each head is also assigned a number which signifies its state .
if it is attached to the track it is assigned @xmath8 and if it is detached from the track it is assigned @xmath9 .
the state @xmath10 corresponds to both the heads of the corresponding side completely detached from the track and that implies the termination of the walk .
so we only consider the states @xmath11 , @xmath12 and @xmath13 .
fig.[trans ] shows the possible transitions and their rates . in order to monitor the motion of one pair of heads on a microtubule and the velocity of this motion , it is simpler and adequate to monitor the motion of the mid - point of the heads @xcite .
the average velocity with which the mid - point moves on the microtubule can be taken as the average velocity with which the microtubule slides past the motor heads .
we define , @xmath14 coordinate of the mid - point , which is identified as follows @xcite : @xmath15 position of the hinge in the @xmath11 state @xmath15 position of he bound head in the @xmath16 state so when both the heads are attached then the mid - point is halfway between the two atpase sites .
as soon as one head detaches , the mid - point shifts to the location of the bound head i.e one of the two binding sites on either side of the previous location of the mid - point .
so from fig.[pair ] we can see that the midpoint jumps @xmath17 and this is the step size of the motion of the mid - point .
a state diagram showing the possible steps of the mid - point and their rates is shown in fig.[step ] .
it is assumed that the binding sites are labelled by half - integer indices and the mid - points between them are , consequently , labelled by integer indices .
so when both the heads are attached , the mid - point is at an integer site and when only one head is attached , the mid - point is at a half - integer site .
so using the idea of peskin and oster explanined above , the motion of a pair of heads on one side can be simplified to the motion of the mid - point of the head alone .
similarly , the same thing can be done for the other pair of heads as well .
so instead of monitoring the motion of four heads we can monitor the motion of the two mid - points like a two - headed motor .
now these two mid - points are coupled together because they are parts of the same 4-headed kinsesin .
this constraint introduces a restriction on the difference between the coordinates of the mid - points .
if @xmath18 and @xmath19 are the coordinates of the mid - points then restriction is @xmath20 .
so this ensures that if the mid - point is at a site @xmath21 on a microtubule then cross - linking can take place only through sites @xmath21 and @xmath22 on the other microtubule .
similarly , if the mid - point on a microtubule is at a site @xmath23 , then the possible cross - linking sites on the other microtubule are @xmath24 and @xmath25 .
the orientations allowed are shown in fig.[orientations ] .
let @xmath26 probabilty of finding the mid - point at @xmath27 at time @xmath28 . from fig.[step ]
we can write the following master equations : @xmath29 @xmath30
now we define @xmath31 and @xmath32 .
@xmath33\end{aligned}\ ] ] @xmath34 similarly , @xmath35\ ] ] @xmath36 in the steady state @xmath37 .
so from equations ( [ m0derivative ] ) and ( [ n0derivative ] ) we get @xmath38 also from normalisation we have , @xmath39 solving equations ( [ ss1 ] ) and ( [ ss2 ] ) , we get @xmath40 mean position = @xmath41 @xmath42\ ] ] @xmath43\ ] ] if we define @xmath44 as the average speed of the mid - point then @xmath45.\ ] ] using the above relation and equations ( [ diffint ] ) and ( [ diffhalfint ] ) we can write @xmath46\\ & & + \sum^{\infty}_{j=-\infty } \left(j+\frac{1}{2}\right)[\beta_bp_j+\beta_fp_{j+1}- { \alpha}p_{j+\frac{1}{2}}]\end{aligned}\ ] ] using @xmath31 and @xmath32 , we get @xmath47 . plugging in the expressions for @xmath48 and @xmath49 from equations ( [ m0 ] ) and ( [ n0 ] ) , we get @xcite @xmath50\label{v}\end{aligned}\ ] ] where , @xmath51 now we couple the motion of the two mid - points which move on their mt tracks with @xmath44 calculated above .
the motion of a mid - point counts only if it is attached to a microtubule and cross - linked to the other microtubule .
so the relative velocity due to the motion of a mid - point on a microtubule is equal to @xmath44 multiplied to the probability of a mid - point being found at any particular site on that microtubule , which is then multiplied by the probability of a mid - point being found at all the corresponding possible cross - linking sites , and then summed over all the sites of the first microtubule .
this prescription can be mathematically expressed as , @xmath52\label{velocity}\end{aligned}\ ] ] here @xmath53 and @xmath54 are the velocities of the mid - points on the first and second microtubules respectively .
similarly , the superscripts on the probabilities also refer to the the microtubules .
but , since there is no functional difference between the two microtubules , these indices can be dropped .
then we have @xmath55 \label{velocity1}\ ] ] we define three new quantities , @xmath56 plugging these expressions in equation ( [ velocity1 ] ) , @xmath57 now , @xmath58 using equation ( [ diffhalfint ] ) @xmath59\ ] ] using the definitions of equations ( [ m ] ) and ( [ n ] ) @xmath60 \label{nderivative}\ ] ] similarly , @xmath61 using equation ( [ diffint ] ) @xmath62\ ] ] using the definitions of equations ( [ m ] ) and ( [ q ] ) @xmath63 \label{qderivative}\ ] ] in the steady state @xmath64 , so we have , @xmath65&=&0 \label{final1}\\ 2[\alpha m + \alpha(1-p)m-(\beta_f+\beta_b)q]&=&0\label{final2}\end{aligned}\ ] ] also since the variables @xmath66,@xmath67 and @xmath68 capture all the cross - linking probabilities ( fig.[orientations ] ) so we can write , @xmath69 solving equations ( [ final1 ] ) , ( [ final2 ] ) and ( [ final3 ] ) we get , @xmath70 using equations ( [ final3 ] ) and ( [ mvalue ] ) , we get the final expression for @xmath71 as , @xmath72 \left(1+\frac{2\alpha(\beta_f+\beta_b)}{2\alpha(\beta_f+\beta_b)+(\beta_f+\beta_b)^2+\alpha^2(2-p)}\right)\label{finalvelocity2}\end{aligned}\ ] ]
from equation ( [ mvalue ] ) it is obvious that @xmath73 .
therefore , @xmath74 which is also consistent with our intuitive expectation . since in our
model cross - linking is allowed through sites next to the adjacent sites on microtubules , hence the two mid - points move with respect to each other .
this factor actually increases the relative velocity above @xmath75 .
the microtubules slide with respect to the mid - points with veocity @xmath44 and the mid - points move with respect to each other which increases the overall relative velocity .
next we consider a few special cases to establish that our results are consitent with physical intuition .
if the probability of the detachment of the back head and the front head is the same and the probability of the free head binding infront and behind the attached head is also the same , then @xmath76 in that case , using equations ( [ sc1 ] ) and ( [ sc2 ] ) in equations ( [ v ] ) and ( [ finalvelocity2 ] ) , we get @xmath77 and @xmath78 respectively , which is expected on physical grounds . if possible orientations for cross - linking are restricted by imposing the condition that cross - linking is allowed only through adjacent sites on two microtubules , then the relative positions of the mid - points with respect to each other do not change
. the relative velocity of sliding will simply be the sum of the velocities with which each mid - point hops on the microtubule .
the velocites are added because of the anti - parallel orientations of the microtubules .
mathematically , in this case @xmath79 and hence , using equation ( [ sc3 ] ) in equation ( [ finalvelocity ] ) we get @xmath80 which agrees with the above prediction .
kruse et al.@xcite calculated the relative velocity of sliding between two filaments due to the action of a two headed motor . in this work the length of the filaments , which is denoted by @xmath81 is assumed to be finite . it is assumed that cross - linking can take place only through adjacent sites of the two microtubules .
a variable @xmath82 is defined , which is the differnce between the coordinates of the minus ends of the microtubules .
thus overlaps can take palce if @xmath83 . the average velocity of sliding as a function of @xmath82 in units of monomers per unit time is @xmath84 where @xmath85 is the probability of cross - linking .
when averaged over all possible values of @xmath82 the final expression for average relative velocity is @xcite @xmath86 now , equation ( [ kruse2 ] ) can be used for our system by taking the limit of @xmath87 and replacing @xmath88 by @xmath75 . in that case , @xmath89 which is consistent with equation ( [ finalvelocity2 ] ) because in the special case cross - linking was allowed only through adjacent sites on two microtubules .
computer simulations as well as analytical studies related to continuum models of motor induced sliding of microtubules and actin filaments have been reported @xcite .
these studies say that two headed bipolar motors slide anti - parallel filaments with a velocity that is twice that of the free motor velocity @xcite . in our case , the motion of kinesin-5 is reduced to a motion of a two headed motor using the model of mid - points . in our case cross - linking
is allowed through adjacent sites on microtubules or sites next to adjacent sites on microtubules . in the continuum
limit the distance between two sites is negligible , so it can be said that all the cross - linking that takes place is effectively through adjacent sites only , as in the continuum limit the adjacent site and the next to adjacent site can be deemed to be the same . hence all the cross - linking can be deemed to be through adjacent sites .
in such a case as we calculated in equation ( [ sc4 ] ) our realtive velocity of sliding @xmath71 is equal to @xmath75 , where @xmath44 is the velocity with which a mid - point slides a filament .
if @xmath44 is taken analogous to the free motor velocity of a two headed motor , then our result in the continuum limit agrees with earlier studies done with continuum models .
i sincerely thank debashish chowdhury for introducing me to molecular motors , and for useful suggestions .
i also thank him for his comments on earlier versions of this manuscript . | kinesin-5 , also known as eg5 in vertebrates is a processive motor with 4 heads , which moves on filamentous tracks called microtubules .
the basic function of kinesin-5 is to slide apart two anti - parallel microtubules by simultaneously walking on both the microtubules @xcite .
we develop an analytical expression for the steady - state relative velocity of this sliding in terms of the rates of attachments and detachments of motor heads with the atpase sites on the microtubules .
we first analyse the motion of one pair of motor heads on one microtubule and then couple it to the motion of the other pair of motor heads of the same motor on the second microtubule to get the relative velocity of sliding . |
the current stage of theoretical cosmology and , in particularly , the dark energy problem calls to consideration of new field theory models which maybe useful for the explanation of the universe acceleration . in the situation
when the attempts to use the well - known classes of field theories as candidates for dark energy ( dark matter ) are not quite successful , such strategy seems to be justified .
moreover , one should bear in mind that number of problems of the early time cosmology ( say , the fundamental cosmological constant problem ) are still too far from the resolution . the ( old or new ) field theoretical model which should include also gravitation itself should solve these problems at once ( at least , in the regime where one can still neglect the quantum gravity effects ) . in the recent paper @xcite the class of dark energy models where dark energy ( scalar or spinor field theory , etc ) couples with some power of the curvature has been proposed . it has been demonstrated on the example of scalar lagrangian coupled with the power of curvature that such modification of gravity - matter coupling term may explain not only the current cosmic speed - up , but also the dark energy dominance assisted by gravity . on the same time
, theories of similar nature have been recently proposed as good candidate for dynamical resolution of the cosmological constant problem @xcite ( see also @xcite ) .
of course , such models are not standard ones , in the sense that they are not multiplicatively renormalizable in curved spacetime @xcite .
hence , they should be considered as kind of effective theories ( without clear understanding of their origin and their relation with more fundamental string / m - theory ) . in the present paper
we continue the study of the properties of such models on the example of scalar self - interacting theory non - linearly coupled with some power of the curvature . admitting the discrete symmetry which prohibits odd powers of scalar in the potential , the effective action formulation ( section 2 )
is developed in close analogy with multiplicatively renormalizable theories @xcite . using riemann normal coordinate expansion ,
the modified green function is evaluated at weak curvature limit ( section 3 ) .
( we work in linear curvature approximation what is expected to be appropriate for the study of quantum effects in the inflationary universe and in late - time , dark energy universe ) .
subsequently , this green function is used for the calculation of the one - loop effective action in the same limit ( section 4 ) .
after the one - loop renormalization of the effective action done in section 5 , one arrives at finite expression which is used for the numerical study of the phase structure in frw universe .
the number of examples explicitly demonstrates the behaviour of the effective potential for such effective theory . in
particularly , it is shown the possibility of curvature - induced phase transitions in the model under discussion ( section 6 ) .
it is indicated that radiative corrections are not very relevant at weak curvature .
finally , in section 7 it is shown that scalar theory under consideration may serve also for dynamical resolution of cosmological constant problem .
the explicit choice of parameters ( which maybe considered as kind of fine - tuning ) giving such dynamical resolution is presented .
it is interesting that radiative corrections are not relative for such a mechanism .
some discussion and outlook are given in the last section .
we extend the @xmath1 theory which is the simplest model where the spontaneous symmetry breaking takes place .
one of the possible couplings of scalar lagrangian with curvature ( at some power ) maybe included into the starting action : @xmath2 . \label{a0}\ ] ] where @xmath3 is an arbitrary mass scale , @xmath4 is the determinant of the metric tensor @xmath5 and @xmath6 is the ordinary lagrangian density of the @xmath1 theory , @xmath7 where @xmath8 is a real scalar field .
our sign convention for the metric is @xmath9 and follows the notation of birrell and davies @xcite .
if the spacetime curvature is constant , the non - linear curvature coupling disappears by the transformation @xmath10 we are interested in a non - trivial effect of the non - linear curvature coupling .
it will be found in a non - static and/or inhomogeneous spacetime .
note that the non - linear coupling of above sort has been proposed in ref .
@xcite ( see also related model in @xcite ) as the model of dark energy which naturally resolves the problem of dark energy dominance in the current universe .
the lagrangian is invariant under the discrete transformation , @xmath11 this symmetry prevents the lagrangian from having @xmath12 terms . a non - vanishing expectation value for the field @xmath13 breaks the discrete symmetry spontaneously .
to evaluate the expectation value for @xmath13 we calculate the effective action .
we start from the generating functional @xmath14 $ ] .
the generating functional of the theory is given by @xmath15 } \equiv \int { \cal d } \phi_0 \ e^{\frac{i}{\hbar}s+\frac{i}{\hbar}\int d^{d}x \sqrt{-g } \phi_0 j } \label{gef}\ ] ] in the presence of the source @xmath16 the classical equation of motion becomes @xmath17 in curved spacetime the functional derivative is defined by @xmath18 - s[\phi_0 ] = \int d^{d}x \sqrt{-g } \frac{\delta s}{\delta \phi_0(x ) } \delta \phi_0(x).\ ] ] we divide the field @xmath8 into a classical background @xmath19 which satisfies the eq.([eom ] ) and a quantum fluctuation @xmath20 , @xmath21 in terms of @xmath19 and @xmath20 the action @xmath22 $ ] is rewritten as @xmath23 & = & s[\phi_{b } + \hbar^{1/2}\tilde{\phi } ] \nonumber \\ & = & s[\phi_{b } ] + \int d^{d}x \sqrt{-g } \left.\frac{\delta s[\phi_0]}{\delta \phi_0(x)}\right|_{\phi_0 = \phi_b}\hbar^{1/2}\tilde{\phi(x ) } \nonumber \\ & & + \frac{\hbar}{2}\int d^{d}x \int d^{d}y \sqrt{-g(x ) } \sqrt{-g(y ) } \tilde{\phi}(x)ig_d^{-1}(x , y)\tilde{\phi}(y ) + \mbox{o}({\tilde{\phi}^3 } ) \nonumber \\ & = & s[\phi_{b } ] -\int d^{d}x \sqrt{-g}j(x ) \hbar^{1/2}\tilde{\phi}(x ) \nonumber \\ & & + \frac{\hbar}{2}\int d^{d}x \int d^{d}y \sqrt{-g(x ) } \sqrt{-g(y ) } \tilde{\phi}(x)ig_d^{-1}(x , y)\tilde{\phi}(y ) \nonumber \\ & & + \mbox{o}({\hbar^{3/2}\tilde{\phi}^3}),\end{aligned}\ ] ] where @xmath24 is given by @xmath25}{\delta \phi_0(x)\delta\phi_0(y ) } \nonumber \\ & = & -\sqrt{-g(x)}\left [ \left(\frac{r}{m^2}\right)^\alpha \left(\square + \xi_0 r - \mu_0 ^ 2 + \frac{\lambda_0}{2}\phi_b^2 \right)\right .
\nonumber \\ & & + \left .
\alpha\left(\frac{r}{m^2}\right)^{\alpha-1 } \frac{r_{;\mu}}{m^2}\nabla^{\mu}\right ] .
\label{gd}\end{aligned}\ ] ] therefore the generating functional ( [ gef ] ) is expanded to be @xmath26 } & = & e^{\frac{i}{\hbar}\left(s[\phi_{b } ] -\int d^{d}x \sqrt{-g}j(x ) \phi_b(x ) \right ) } \nonumber \\ & & \times \int { \cal d } \tilde{\phi}\ e^{\frac{i}{2 } \int d^{d}x \int d^{d}y \sqrt{-g(x ) } \sqrt{-g(y ) } \tilde{\phi}(x)ig_d^{-1}(x , y)\tilde{\phi}(y)+\mbox{o}(\hbar^{1/2})}\end{aligned}\ ] ] performing the path - integral , @xmath27^{-1/2 } , \ ] ] we obtain the generating functional @xmath28 = s[\phi_{b } ] -\int d^{d}x \sqrt{-g}j(x ) \phi_b(x ) -\frac{i\hbar}{2}\ln \left[\mbox{det}\ ig_d^{-1}\right ] + \mbox{o}(\hbar^{3/2 } ) .
\label{gf0}\ ] ] the effective action @xmath29 $ ] is given by the legendre transform of @xmath14 $ ] , @xmath30 \equiv w[j ] - \int d^{d}x \sqrt{-g } \phi_c(x)j(x ) , \label{ea0}\ ] ] where @xmath31 denotes the expectation value of @xmath8 in the presence of the source @xmath16 , @xmath32}{\delta j}.\ ] ] substituting eq .
( [ gf0 ] ) to eq .
( [ ea0 ] ) we find that the effective action is expanded to be @xmath30 = s[\phi_c ] -\frac{i\hbar}{2}\ln \left[\mbox{det}\ ig_d^{-1}\right ] + \mbox{o}(\hbar^{3/2 } ) .
\label{ea}\ ] ] in the path integral formalism the expectation value of @xmath8 in the ground state satisfies @xmath33}{\delta j}.\ ] ] this equation is rewritten as @xmath34}{\delta \phi_c } \right|_{\phi_c \rightarrow \langle \phi_0 \rangle } = 0 , \label{gap : action}\ ] ] where we have used the relation @xmath35}{\delta \phi_c } = j.\ ] ] the equation ( [ gap : action ] ) is called the gap equation .
the solution of the gap equation shows the presence of stationary points of the effective action @xmath36 $ ] .
thus , the phase structure of the theory maybe found by observing the stationary points of @xmath36 $ ] ( for a review of the quantum field theory in curved spacetime , see ref .
following the procedure developed in refs .
@xcite , we solve the klein - gordon equation and calculate the green function @xmath37 ( [ gd ] ) at the weak curvature limit . note that in green function calculation we actually work in the expansion on the curvatures what is known to be appropriate for the study of quantum effects in the inflationary universe and also in late - time , dark energy universe .
the green function @xmath37 satisfies the modified klein - gordon equation , @xmath38ig_d(x , x')=\delta^d(x - x ' ) .
\label{kg}\end{aligned}\ ] ] for a practical calculations it is more convenient to introduce @xmath39 and @xmath40 which are defined by @xmath41 and @xmath42 ( -g)^{-1/4}(x)\left(\frac{r(x)}{m^2}\right)^{-\alpha/2 } , \label{ham00}\end{aligned}\ ] ] in terms of @xmath39 and @xmath40 the klein - gordon equation ( [ kg ] ) reads @xmath43 after some calculations the hamiltonian density @xmath40 is rewritten as @xmath44 where the potential @xmath45 is given by @xmath46 + \xi_0 r - \mu_0 ^ 2 + \frac{\lambda_0}{2}\phi^2 \nonumber \\ & & + \alpha \frac{r_{;\mu}}{r } \left[(-g)^{-1/4}\partial^\mu(-g)^{1/4 } \right ] -\frac{\alpha}{2}\left(\frac{\alpha}{2}-1\right ) \frac{r^{;\mu}r_{;\mu}}{r^2 } -\frac{\alpha}{2}\frac{\square r}{r } .\end{aligned}\ ] ] here we adopt the riemann normal coordinate expansion @xcite .
we set the origin of the coordinates at @xmath47 and write @xmath48 and neglect terms proportional to @xmath49 , @xmath50 .
hence , the eq.([kg2 ] ) is rewritten as @xmath51 where @xmath52\right|_{y=0 } \nonumber \\ & \equiv & -\mu_0 ^ 2 + u , \\
\beta_\mu & = & \left .
\left ( u_{;\mu } + \frac{\alpha}{6}\frac{r^{;\nu}}{r}r_{\mu\nu } \right)\right|_{y=0}.\end{aligned}\ ] ] performing the fourier transformation @xmath53 one finds that the equation ( [ kg3 ] ) becomes @xmath54 first we consider the solution for a positive @xmath55 . in this case
the following proper - time form for the solution of eq.([kg4 ] ) is assumed @xmath56 substituting this into the eq.([kg4 ] ) one obtains @xmath57 where @xmath58 on the other hand , there is an identity @xmath59 comparing the eq.([as2 ] ) with the eq.([as3 ] ) , the simultaneous differential equations are obtained @xmath60 the solutions of these differential equations are found to be @xmath61 where @xmath62 , @xmath63 and @xmath64 are constant matrix parameters which are determined by the boundary conditions . in the flat space time , @xmath65 , eq.([kg4 ] ) must reproduce a klein - gordon equation in the minkowski space .
thus , the functions @xmath66 , @xmath67 and @xmath68 must satisfy @xmath69 at the limit @xmath70 , @xmath71 , @xmath72 .
these boundary conditions fix the parameters @xmath62 , @xmath63 and @xmath64 to be @xmath73 therefore the functions @xmath66 , @xmath67 and @xmath68 are simplified to @xmath74.\end{aligned}\ ] ] since the parameters @xmath70 , @xmath71 and @xmath75 are small enough at the weak curvature limit , we neglect the higher order term about these parameters for a small @xmath76 and get @xmath77 inserting these solutions into eq.([ass ] ) , the green function @xmath78 follows @xmath79 where a scale @xmath80 is introduced .
it corresponds to the upper limit where we can neglect the higher order term about the parameters @xmath70 , @xmath71 and @xmath75 . since the integrand becomes smaller and smaller above the scale @xmath81 , we drop the contribution from the region above @xmath81 .
it is straightforward to extend above analysis to a negative @xmath55 and obtain @xmath82 integrating over @xmath76 in eq.([gpp ] ) and ( [ gpp2 ] ) , it follows @xmath83 \nonumber \\ & & \times \left[1+\mbox{erf}\left(-\mbox{\scriptsize sgn}(\chi^2)\frac{p^2-\chi^2-i\zeta_\mu p^\mu } { \sqrt{4(p^{\mu}\xi_{\mu\nu}p^{\nu}-i\beta_\mu p^\mu)}}\right)\right ] , \end{aligned}\ ] ] where erf@xmath84 is the error function . in the configuration space
the green function @xmath85 reads @xmath86 at the origin it reduces to @xmath87 to perform the momentum integral the wick rotation maybe applied . neglecting the higher order term about @xmath70 , @xmath71 and @xmath75
, we obtain @xmath88 hence , the scalar green function is found at small curvature .
it will be used in the calculation of the one - loop effective action .
the effective action of the present model is given by ( [ ea ] ) .
here we normalize the effective action to satisfy @xmath89 = s[\phi = 0].\ ] ] in this normalization the one - loop effective action is given by @xmath90 = s[\phi ] -\frac{i\hbar}{2}\left[\ln \left\{\mbox{det}\ ig_d^{-1}(\phi^2)\right\ } -\ln \left\{\mbox{det}\ ig_d^{-1}(\phi^2 = 0)\right\}\right ] .
\label{ea2}\ ] ] the right hand side of eq .
( [ ea2 ] ) contains a term proportional to @xmath91 from the eq.([g0 ] ) the green function at the weak curvature limit is found to be @xmath92 the relationship between @xmath37 and @xmath93 is given by @xmath94 the effective action at the weak curvature limit reduces to @xmath95 & = & s[\phi ] -\frac{\hbar\lambda_0}{4(4\pi)^{d/2 } } \\
\nonumber & & \times \int d^d x \sqrt{-g}\int^{\phi^2}_{0 } dm^2 \int^{\mbox{\scriptsize sgn}(\chi^2)\lambda}_{0 } \frac{ds}{s^{d/2 } } \left ( 1+\frac{1}{2 } \mbox{tr } \xi s \right ) e^{-\chi^2(\phi^2=m^2 ) s } .
\label{ea3}\end{aligned}\ ] ] integrating over @xmath96 one gets @xmath95 & = & s[\phi ] + \frac{\hbar}{2(4\pi)^{d/2 } } \int d^d x \sqrt{-g } \nonumber \\ & & \times \int^{\mbox{\scriptsize sgn}(\chi^2)\lambda}_{0 } \frac{ds}{s^{d/2 + 1 } } \left ( 1+\frac{1}{2 } \mbox{tr } \xi s \right ) \left(e^{-\chi^2(\phi^2 ) s}-e^{-\chi^2(\phi^2=0 ) s } \right ) .
\label{ea4p}\end{aligned}\ ] ] performing the integration about @xmath76 and neglecting the higher order term about @xmath97 , we get @xmath95 & = & s[\phi ] \nonumber \\ & & + \frac{\hbar}{2(4\pi)^{d/2}}\gamma\left(-\frac{d}{2}\right ) \int d^d x \sqrt{-g}\left [ ( \chi^2(\phi^2))^{d/2 } \left ( 1-\frac{d \mbox{tr } \xi}{4\chi^2(\phi^2)}\right)\right .
\nonumber \\ & & \left .
\hspace*{2 cm } - ( \chi^2(\phi^2=0))^{d/2 } \left ( 1-\frac{d \mbox{tr } \xi}{4\chi^2(\phi^2=0)}\right ) \right ] .
\label{ea4}\end{aligned}\ ] ] the effective lagrangian density is defined by @xmath98=\int d^d x \sqrt{-g } { \cal l}_{eff}.\ ] ] from eqs.([a0 ] ) and ( [ ea4 ] ) it is found to be @xmath99 .
\label{al}\end{aligned}\ ] ] with @xmath100 in the limit @xmath101 and @xmath102 the effective lagrangian density ( [ al ] ) reproduces the one in the minkowski space .
the effective action ( [ ea4 ] ) and the effective lagrangian density ( [ al ] ) are divergent in two and four dimensions . to obtain the finite value we must renormalize the theory .
( remind that theory is not multiplicatively renormalizable in curved spacetime @xcite ) . at the present order
we impose the following renormalization conditions @xmath103}{\partial \phi^2 } \right|_{\phi = m_1 } \equiv \left(\frac{r}{m^2}\right)^\alpha
( \mu_r^2 - \xi_r r ) , \label{def : mur}\ ] ] @xmath104}{\partial \phi^4 } \right|_{\phi = m_2 } \equiv -\left(\frac{r}{m^2}\right)^\alpha \lambda_r .
\label{def : lr}\ ] ] where @xmath105 and @xmath106 are the renormalization scales . and the cosmological constant . ] from these conditions one obtains the relationship between the renormalized parameters @xmath107 , @xmath108 , @xmath109 and the bare ones @xmath110 , @xmath111 , @xmath112 .
@xmath113 \right .
\nonumber\\ & & \left .
+ \frac{1}{2}\gamma\left(2-\frac{d}{2}\right ) ( \chi^2(m_1 ^ 2))^{d/2 - 2 } \right . \nonumber \\ & & \left.\hspace*{18mm}\times \left [ 1+\lambda_0 m_1 ^ 2\left(\frac{d}{2}-2\right)(\chi^2(m_1 ^ 2))^{-1}\right ] \mbox{tr}\xi \right\ } \nonumber \\ & \equiv & \mu_0 ^ 2 - \xi_0 r -\frac{\lambda_0}{2}m_1 ^ 2 + \frac{\hbar}{2(4\pi)^{d/2}}\left(\frac{r}{m^2}\right)^{-\alpha } \lambda_0 f(d , m_1),\end{aligned}\ ] ] @xmath114 \right . \nonumber\\ & & \left .
+ \frac{1}{2}\gamma\left(3-\frac{d}{2}\right ) ( \chi^2(m_2 ^ 2))^{d/2 - 3}\left [ 3 + 6\lambda_0 m_2 ^ 2\left(\frac{d}{2}-3\right)(\chi^2(m_2 ^ 2))^{-1 } \right.\right . \nonumber \\ & & \left.\left.\hspace*{28 mm } + \lambda_0 ^ 2 m_2 ^ 4\left(\frac{d}{2}-3\right)\left(\frac{d}{2}-4\right ) ( \chi^2(m_2 ^ 2))^{-2 } \right ] \mbox{tr}\xi \right\ } \nonumber \\ & \equiv & \lambda_0 + \frac{\hbar}{2(4\pi)^{d/2}}\left(\frac{r}{m^2}\right)^{-\alpha } \lambda_0 ^ 2 g(d , m_2 ) .\end{aligned}\ ] ] by using these renormalized parameters the effective lagrangian density ( [ al ] ) is rewritten as @xmath115 .
\label{alr}\end{aligned}\ ] ] at the two - dimensional limit eq.([alr ] ) becomes @xmath116 .\end{aligned}\ ] ] taking the four - dimensional limit of eq.([alr ] ) , we obtain @xmath117 . \nonumber\end{aligned}\ ] ] therefore the ultra - violet divergences in the effective lagrangian density ( [ al ] ) and the effective action ( [ ea4 ] ) disappear after the one - loop renormalization .
note that to make sense to divergent expressions one can apply any other regularization ( say , zeta - regularization @xcite ) . of course , the question of dependence from the regularization in such effective models remains .
it is expected that the non - linear curvature coupling in our model leads to non - trivial consequences for the phase structure in a non - static spacetime .
( this may have the interesting applications for early - time , accelerating and late - time , accelerating universe when quantum gravity effects are negligible ) . here
we consider the model in the @xmath118-dimensional fridmann - robertson - walker spacetime . by spirit
it is similar to the study of phase structure for scalar self - interacting theory , scalar electrodynamics and guts in curved spacetime ( see @xcite and for complete list of refs . and review @xcite ) where curvature - induced phase transitions were discovered .
the frw spacetime is defined by the metric @xmath119 , \ ] ] where @xmath120 corresponds to the curvature of the @xmath121-dimensional space .
the curvature @xmath122 in this spacetime is given by @xmath123 where the hubble rate @xmath124 is defined by @xmath125 .
the first derivative of the curvature is @xmath126 , \\
r_{;i } & = & 0 .\end{aligned}\ ] ] thus , it follows @xmath127 ^ 2 , \label{rr1 } \\
\square r & = & 2(d-1 ) \left[\dddot{h } + ( 2d-1)\ddot{h}h + d(d-1)\dot{h}h^2+d\dot{h}^2 \right .
\nonumber \\ & & \left .
-(d-2)\left ( d h^2+\dot{h}-3\frac{h}{a}\right)\frac{k}{a^2}\right ] .
\label{rr2}\end{aligned}\ ] ] here the spacetime with @xmath128 and @xmath129 is considered . in this spacetime
the curvature invariants are @xmath130 \frac{1}{t^2}.\end{aligned}\ ] ] the ground state of the theory depends on the spacetime structure and the parameters of the theory under consideration .
to study the phase structure of the @xmath0 theory in frw spacetime we evaluate the effective action , @xmath131 .
the expectation value of the field @xmath13 is obtained by searching of the stationary points of the effective action .
the effective action develops a small imaginary part for a negative @xmath132 .
we evaluate a real part of the effective action to find the ground state . in an expanding universe
the scale factor @xmath133 increases as the time runs .
here we suppose that the scale factor @xmath133 is proportional to the time @xmath134 , i.e. @xmath135 and numerically calculate the renormalized effective action ( [ alr ] ) .
all mass scales are normalized by an arbitrary mass scale @xmath3 and @xmath136 .
the renormalization scale is chosen to be @xmath137 and @xmath138 .
first we consider the stationary and spatially homogeneous @xmath13 . in this case
the kinetic term of @xmath13 disappears .
one can define the ordinary effective potential @xmath139 by @xmath140 the expectation value of the field @xmath13 is defined by the minimum of the effective potential .
behaviour of the effective potential and the mass gap for @xmath141 , @xmath142 , @xmath143 , @xmath144 , @xmath145 and @xmath146 . ]
behaviour of the effective potential and the mass gap for @xmath141 , @xmath142 , @xmath143 , @xmath144 , @xmath145 and @xmath146 . ]
behaviour of the effective potential and the mass gap for @xmath147 , @xmath142 , @xmath143 , @xmath144 , @xmath145 and @xmath146 . ]
behaviour of the effective potential and the mass gap for @xmath147 , @xmath142 , @xmath143 , @xmath144 , @xmath145 and @xmath146 . ]
behaviour of the effective potential and the mass gap for @xmath148 , @xmath142 , @xmath143 , @xmath144 , @xmath145 and @xmath146 . ]
behaviour of the effective potential and the mass gap for @xmath148 , @xmath142 , @xmath143 , @xmath144 , @xmath145 and @xmath146 . ]
we illustrate the behavior of the effective potential for a minimal gravitational coupling @xmath143 at @xmath144 in figs.1 , 2 and 3 . here
we set @xmath146 to regularize the theory with keeping the general covariance .
it corresponds to the cut - off scale , @xmath149 , where @xmath106 is a renormalization scale .
since the spacetime curvature behaves as @xmath150 , it raises up for a small @xmath134 .
it is not valid to apply the weak curvature expansion at the beginning of the universe . in figs . 1 , 2 and 3 we solve the gap equation only for @xmath151 .
as is shown in the figures the effective potential seems to be almost proportional to @xmath152 . in figs . 1 and 2 the expectation value @xmath153 is located near the stationary point of the classical potential , @xmath154 . for the minimal gravitational coupling
the classical part of the effective action has no curvature dependence except for the overall factor and the einstein - hilbert term .
on the other side , the radiative correction depends on the curvature . at @xmath155
our model reduces to the ordinary @xmath0 theory .
as is known , the radiative correction has only a small contribution to spontaneous symmetry breaking for a large @xmath107 in the ordinary theory . in the case of a negative @xmath156 the classical part of the effective action ( [ alr ] )
is enhanced as the curvature decreases .
contributions from the radiative corrections are lost at a large @xmath134 . for a positive @xmath156 the classical part of the effective action
is suppressed except for the einstein - hilbert term as the curvature decreases .
hence we observe the curvature dependence of the expectation value @xmath153 only if @xmath157 , see fig .
behaviour of the effective potential and the mass gap for @xmath141 , @xmath142 , @xmath143 , @xmath158 , @xmath145 and @xmath146 . ]
behaviour of the effective potential and the mass gap for @xmath141 , @xmath142 , @xmath143 , @xmath158 , @xmath145 and @xmath146 . ]
behaviour of the effective potential and the mass gap for @xmath141 , @xmath142 , @xmath159 , @xmath158 , @xmath145 and @xmath146 . ] behaviour of the effective potential and the mass gap for @xmath141 , @xmath142 , @xmath159 , @xmath158 , @xmath145 and @xmath146 . ]
next we plot the behaviour of the effective potential for @xmath158 in figs . 4 , 5 , 6 and 7 . here
we consider the minimal and conformal gravitational coupling , @xmath143 and @xmath160 $ ] respectively .
( of course , the scalar sector is conformally invariant for correspondent choice of @xmath71 only when @xmath141 . ) for @xmath158 spontaneous symmetry breaking can not take place on the classical level of the ordinary @xmath0 theory .
the radiative correction has an essential role to break the symmetry in the ordinary theory , i.e. @xmath141 .
it is known as a radiative symmetry breaking .
first order phase transition is observed in figs . 4 and 5 .
however , the radiative correction has only a small effect .
the expectation value generated by the radiative symmetry breaking is extremely small , as is shown in figs . 4 and 5 at @xmath141 .
a long time is necessary until phase transition takes place .
behaviour of the effective potential and the mass gap for @xmath148 , @xmath142 , @xmath143 , @xmath158 , @xmath145 and @xmath146 . ]
behaviour of the effective potential and the mass gap for @xmath148 , @xmath142 , @xmath143 , @xmath158 , @xmath145 and @xmath146 . ]
behaviour of the effective potential and the mass gap for @xmath148 , @xmath142 , @xmath159 , @xmath158 , @xmath145 and @xmath146 . ]
behaviour of the effective potential and the mass gap for @xmath148 , @xmath142 , @xmath159 , @xmath158 , @xmath145 and @xmath146 . ] in the case of a positive @xmath156 the ratio of the radiative correction to the classical part of the @xmath0 lagrangian is enhanced as the curvature decreases . in figs . 6 and 7
we also observe the first order phase transition and find that the expectation value @xmath153 is extremely enhanced in comparison with the case @xmath141 .
for @xmath147 only a symmetric phase is realized at @xmath158 and then radiative symmetry breaking does not occur . as is shown in figs . 6 and 7
, the expectation value of @xmath13 has non - negligible time dependence for the case @xmath161 .
hence we next consider the spacialy homogeneous but time dependent @xmath13 at @xmath161 and @xmath162
. then we should solve the gap equation ( [ gap : action ] ) . from eq .
( [ gap : action ] ) the equation of motion for @xmath13 is obtained @xmath163 = 0 . \label{eqm}\end{aligned}\ ] ] to find an exact solution one needs to solve this equation for a general time - dependent form of @xmath13 .
however , it is instructive to consider the solution of eq.([eqm ] ) for a special form of @xmath13 . in the present paper
it is assumed that @xmath164 where @xmath165 is a constant parameter .
here we fix the parameter @xmath47 at @xmath166 or @xmath167 and numerically solve the equation of motion for @xmath13 . under the assumption ( [ ass2 ] )
eq.([eqm ] ) reads @xmath168 \frac{x}{t^2}\phi(t ) + \left(\frac{r}{m^2 } \right)^{-\alpha } \frac{\partial v(\phi)}{\partial \phi } = 0 .
\label{eqm2}\ ] ] thus the expectation value @xmath153 is found by observing the stationary point of the following function @xmath169\equiv\left[(x-1)+(d-1)h_0 - 2\alpha\right ] \frac{x}{2 t^2}\phi(t)^2 + \left(\frac{r}{m^2 } \right)^{-\alpha } v(\phi ) .
\label{eff2}\ ] ] the solution is shown in figs . 8 and 9 . as is seen in figs .
8 and 9 , we observe the first order phase transition again .
in the case of the minimal gravitational coupling we find the two steps of the first order phase transition for @xmath170 . in fig . 8 ( a ) the solution becomes unstable for @xmath171 .
the weak curvature approximation and/or the assumption @xmath172 can not be applicable for such a small @xmath134 . in all the cases
the mass scale @xmath165 depends on the time @xmath134 .
however , we observe that the mass scale @xmath165 is almost static between @xmath173 and @xmath174 for @xmath170 and between @xmath174 and @xmath175 for @xmath176 .
it is expected that there is a solution with gradually decreasing @xmath47 after the first order phase transition near the time @xmath177 . on the other hand
, we observe in figs .
6 , 7 , 8 and 9 that the time of the first order transition near @xmath177 has no large dependence on @xmath47 in the present analysis at @xmath148 and @xmath142 .
it seems to be a characteristic feature of the theory with @xmath148 in the frw spacetime .
behaviour of the mass scale @xmath165 for @xmath148 , @xmath142 , @xmath143 , @xmath158 , @xmath145 and @xmath146.,title="fig : " ] + ( a ) @xmath178 behaviour of the mass scale @xmath165 for @xmath148 , @xmath142 , @xmath143 , @xmath158 , @xmath145 and @xmath146.,title="fig : " ] + ( b ) @xmath179 behaviour of the mass scale @xmath165 for @xmath148 , @xmath142 , @xmath159 , @xmath158 , @xmath180 and @xmath146.,title="fig : " ] + ( a ) @xmath178 behaviour of the mass scale @xmath165 for @xmath148 , @xmath142 , @xmath159 , @xmath158 , @xmath180 and @xmath146.,title="fig : " ] + ( b ) @xmath179 hence , in the present section we numerically estimated phase structure for scalar theory in curved spacetime under consideration .
the phase structure strongly depends on the sign of @xmath156 .
compared with the usual @xmath0 theory in weakly curved space , i.e. @xmath141 , a positive @xmath156 significantly enhances the symmetry breaking . for a negative @xmath156 the radiative symmetry breaking does not take place .
as it was shown , mainly first - order curvature induced phase transitions may occur here . in all figures
we put @xmath146 and @xmath142 .
the spacetime curvature is given by eq.([r : frw ] ) and found to be @xmath181 .
the typical scale found to be @xmath182 , @xmath183 in figs .
1@xmath1843 , 6@xmath1849 and @xmath185 , @xmath186 in figs . 4 , 5 .
the curvature scales for such time scales are given by @xmath187 and @xmath188 respectively .
if we consider the electroweak scale higgs boson , @xmath189gev@xmath190 , we obtain the typical mass scale @xmath191gev in the former case and @xmath192gev in the latter case .
the cut - off scales which corresponds to @xmath146 are @xmath193tev , near the scale of a large extra dimension , and @xmath194gev , respectively .
for both cases the typical scale of time and curvature are given by @xmath195gev@xmath196 and @xmath197gev@xmath198 .
thus the quantum correction induced by the curvature effect has an interesting contribution to the effective action at the very early stage of the universe .
the time scale of the phase transition depends on the coupling constant @xmath199 . for a stronger coupling
the radiative correction is enhanced and a smaller time scale is obtained .
thus we consider the regularized theory and calculate the effective potential in a d - dimensional spacetime with @xmath146 .
as is shown in the previous section , we obtain the finite effective potential ( [ alrd4 ] ) in four spacetime dimensions . here
we show some results in four dimensions . evaluating the effective potential ( [ alrd4 ] ) we draw the solution of the gap equation .
the first order phase transitions is observed as is shown in fig . 10 .
compared with the case @xmath146 ( figs . 6 and 7 ) the critical time becomes late and the the two steps transition does not realised in four dimensions .
behaviour of the effective potential and the mass gap for @xmath148 , @xmath142 , @xmath158 , @xmath200 in @xmath201.,title="fig : " ] + ( a ) @xmath143 behaviour of the effective potential and the mass gap for @xmath148 , @xmath142 , @xmath158 , @xmath200 in @xmath201.,title="fig : " ] + ( b ) @xmath159 the main lesson of this consideration is that it is unlikely that radiative corrections may strongly influence the dark energy universe in such theory .
however , quantum effects may become quite important for phantom universe where big rip may occur @xcite .
it was already demonstrated that in such circumstances already quantum effects of free fields may prevent big rip occurrence @xcite .
however , in order to analyse this problem in our framework one needs the calculation of the one - loop effective action at strong curvature .
there are some proposals to solve the cosmological constant problem dynamically @xcite .
one of the possible solutions of the cosmological constant problem is pointed out by mukohyama and randall @xcite . in @xcite ,
the following type of the action similar to the one under consideration has been proposed : @xmath202\ , \ ] ] where @xmath203 is a proper function of the scalar curvature @xmath122 .
when the curvature is small , it is assumed @xmath203 behaves as @xmath204 here @xmath205 is positive .
when the curvature is small , the vacuum energy , and therefore the value of the potential becomes small .
then one may assume , for the small curvature , @xmath206 behaves as @xmath207 here @xmath208 and @xmath209 are constants . if @xmath210 , the factor in front of the kinetic term of @xmath211 in ( [ mr1 ] ) becomes large .
this makes the time development of the scalar field @xmath211 very slow and it is expected that @xmath211 does not reach @xmath209 .
this model may explain the acceleration of the present universe .
the model ( [ mr1 ] ) is also expected to be stable under the radiative corrections .
in fact , when the curvature is small and the time development of the curvature can be neglected , if we rescale the scalar field @xmath211 as @xmath212 , the curvature in the kinetic term can be absorbed into the redefinition of @xmath211 and there appear factors including @xmath213 .
therefore if @xmath214 , the interactions could be suppressed when the curvature is small and there will not appear the radiative correction to the vacuum energy except the one loop corrections .
we now show that there is an exactly solvable model , which realizes the scenario in @xcite .
assume @xmath215 and ( [ mrio1 ] ) and ( [ mrio2 ] ) are exact : @xmath216 here @xmath75 is a constant .
@xmath217 term is neglected by putting @xmath218 in ( [ mr1 ] ) since the curvature is small . then by the variation over the metric , one obtains @xmath219 on the other hand , the variation over @xmath211 gives @xmath220 we assume the four dimensional frw metric with flat spatial part : @xmath221 it is chosen that @xmath211 only depends on the time coordinate @xmath134 .
then the @xmath222-component of ( [ mrio4 ] ) gives @xmath223 where @xmath124 is the hubble rate . on the other hand , ( [ mrio6 ] )
gives @xmath224 a solution of eqs.([mrio7 ] ) and ( [ mrio8 ] ) is given by @xmath225 when @xmath226 . in
case @xmath227 , this scale factor @xmath228 ( [ mrio9 ] ) does not express expanding universe but shrinking one .
if the direction of time is changed as @xmath229 , the expanding universe emerges with scale factor @xmath230 . in the expression , however , since @xmath231 is not always an integer , @xmath134 should be negative so that the scale factor should be real . to avoid this apparent difficulty , we may further shift the origin of the time as @xmath232 .
then the time @xmath134 can be positive as long as @xmath233 .
when @xmath227 , instead of ( [ mrio9 ] ) , one may propose @xmath234 the assumption ( [ mrio9 ] ) or ( [ mrio9b ] ) reduces eqs.([mrio7 ] ) and ( [ mrio8 ] ) to the algebraic equations : @xmath235 which gives @xmath236 since @xmath237 should be positive , one finds @xmath238 by properly choosing the parameter @xmath75 and @xmath208 , we can always obtain a solution as in ( [ mrio9 ] ) or ( [ mrio9b ] ) .
for example , if @xmath239 which gives the following value of the parameter @xmath240 of the equation of state @xmath241 we find @xmath242 it is interesting that the value of @xmath240 ( [ aa2 ] ) is consistent with the observed one . for @xmath226 case , corresponding to ( [ mrio9 ] ) , since @xmath243 , the curvature @xmath122 decreases as @xmath150 with time @xmath134 .
eq.([mrio9 ] ) tells that @xmath211 approaches to @xmath209 but does not arrive at @xmath209 in a finite time , as expected in @xcite . as seen from the expression of @xmath124 in ( [ mrio9 ] ) or ( [ mrio9b ] ) , if we substitute the value of the age of the present universe @xmath244years@xmath245ev@xmath246 into @xmath134 or @xmath247 , the observed value of @xmath124 could be reproduced , which could explain the smallness of the effective cosmological constant @xmath248 . note that even if there is no potential term , that is , @xmath249 , when @xmath250 , there is a solution @xmath251 which gives the parameter @xmath240 of the equation of state : @xmath252 , although @xmath240 is not realistic .
here we consider the solution of the cosmological constant problem in the theory ( [ a0 ] ) . as is shown in the previous section
, one could expect that the radiative correction is suppressed when the curvature is small if @xmath253 .
as in the case of ( [ mr1 ] ) , if we rescale @xmath13 by @xmath254 , the kinetic term becomes standardly normalized one and does not include the curvature .
the @xmath0 interaction terms , however , include the factor @xmath255 and may be suppressed for the small curvature .
hence , there would not appear the radiative correction to the vacuum energy except the one - loop corrections .
as an example of the solvable case , we put @xmath147 and @xmath256 .
the variation over the metric gives @xmath257g_{\mu\nu } + \frac{r_{\mu\nu}}{2\kappa^2 } \nonumber \\ & & + \frac{m^2}{2r}\partial_\mu \phi \partial_\nu \phi - \frac{m^2 r_{\mu\nu}}{r^2}\left\ { \frac{1}{2}\partial_\rho \phi \partial^\rho \phi + \frac{\lambda}{4!}\phi^4\right\ } \nonumber \\ & & - \left(\nabla_\mu \nabla_\nu - g_{\mu\nu } \nabla^2\right)\left\{\frac{m^2}{r^2}\left ( \frac{1}{2}\partial_\rho \phi \partial^\rho \phi + \frac{\lambda}{4!}\phi^4\right)\right\}\ .\end{aligned}\ ] ] scalar field equation is @xmath258 by assuming the frw metric ( [ mr11 ] ) and @xmath259 , eqs.([mrio15 ] ) and ( [ mrio16 ] ) have the following forms : @xmath260 with the anzats @xmath261 when @xmath226 or @xmath262 when @xmath227 .
eqs.([mrio17 ] ) and ( [ mrio18 ] ) reduce to the algebraic equations : @xmath263 which give @xmath264 therefore , with the proper choice of parameters , the solution in the form of ( [ mrio19 ] ) or ( [ mrio19b ] ) follows .
we should note the behavior of @xmath124 ( [ mrio19 ] ) or ( [ mrio19b ] ) is almost identical with that in ( [ mrio9 ] ) or ( [ mrio9b ] ) .
one may consider the model including two scalar fields @xmath211 and @xmath13 as @xmath265\ .\end{aligned}\ ] ] even in this case , for frw metric ( [ mr11 ] ) and assuming ( [ mrio9 ] ) or ( [ mrio9b ] ) and/or ( [ mrio19 ] ) or ( [ mrio19b ] ) , we obtain ( [ mrio18 ] ) and ( [ mrio21 ] )
. the equation corresponding to ( [ mrio17 ] ) and/or ( [ mrio20 ] ) has the following form : @xmath266 then by properly choosing the parameters , we may obtain an exact solution for cosmological constant again .
hence , we demonstrated that the mechanism proposed in ref.@xcite to dynamically solve the cosmological constant problem is naturally realized also for the class of models investigated in this work .
the final remark is in order . in @xcite
( see also @xcite ) , the cosmological constant damping in the model contained @xmath267 term in the lagrangian has been investigated .
it has been shown the possibility to solve the cosmological constant problem .
although the problem could be solved in the model , the effective newton constant @xmath268 suffers the correction like @xmath269 with a constant @xmath270 . here
@xmath271 is the bare newton constant . in the solution of model @xcite
, @xmath13 behaves as a linear function of the time @xmath134 .
naturally @xmath134 could be the age of the present universe , @xmath272 , the correction becomes very large and the effective newton constant becomes unnaturally small . in the model under consideration ( [ a0 ] ) with ( [ i2 ] ) , similar term is included but due to the factor @xmath273 , if @xmath274 , the correction is suppressed since @xmath275 and @xmath122 is very small . in general @xmath156 can be also negative . in the solvable case in ( [ mrio15 ] ) , the correction could be @xmath276 .
since the @xmath277 or @xmath278 , by combining ( [ mrio19 ] ) or ( [ mrio19b ] ) , the correction takes the order of unity .
then the effective newton constant could receive the correction but it is finite , which does not lead to the instabilities of the above sort ( smallness of gravitational constant ) .
in summary , we studied the scalar self - interacting theory whose lagrangian non - linearly couples with some power of the curvature as the effective field theory .
the one - loop effective action at weak curvature is found here using specific regularization scheme .
the phase structure of the theory is carefully investigated , the examples of curvature - induced phase transitions are presented numerically ( for the case of positive power of curvature ) .
the comparison with the ordinary scalar self - interacting theory in curved space is done . for negative value of @xmath156 the radiative symmetry breaking
does not take place .
it remains to develop the same one - loop formulation at strong curvature as in such a case the phase transitions maybe considered in the vicinity of big rip . the dynamical mechanism to explain the origin of the extremely small cosmological constant in the spirit of papers @xcite is presented .
it is shown that in the class of the models under consideration ( including the case with two scalars ) such a proposal maybe realized quite successfully . unfortunately , despite the preliminary expectations the radiative corrections for the theory under consideration are not relevant in the study of the dynamical cosmological constant .
there are many directions where our approach maybe generalised . in
particularly , it is known that phase structure of njl model in curved spacetime is quite rich ( for a review , see @xcite ) .
hence , it would be extremely interesting to consider the njl - like models non - linearly coupled with power of the curvature in the same spirit .
such models maybe even more attractive due to fact that @xmath279 expansion is more reliable than one - loop approximation in non - renormalizable theories .
one can also consider yang - mills theory ( or born - infeld theory ) coupled to curvature powers in the same way .
definitely , this may bring the number of the interesting applications in such issues as dark energy and cosmological constant . from another point ,
the new matter - gravity coupling @xcite maybe also considered as kind of modification of gravitation itself .
in such a situation , when theory somehow deviates from general relativity , it is not clear apriori if the metric formalism and first - order formulation should lead to the same physical picture @xcite .
( indeed , unlike to general relativity where metric and palatini formulations basically coincide , it is not the case for our theory ( 1)@xcite already on classical level @xcite )
. it would be extremely interesting to develop the one - loop effective action formulation in the same way as the one developed in this paper when an external gravity is formulated in palatini form .
this will be done elsewhere .
this research has been supported in part by jsps grant - in - aid for scientific research , no .
13135208 ( s.n . ) and in part by rfbr grant 03 - 01 - 00105 and lrss grant 1252.2003.2 ( s.d.o . )
99 nojiri s and odintsov s d 2004 b * 599 * 137 dolgov a d and kawasaki m 2003 _ preprint _ astro - ph/0307442 mukohyama s and randall l 2004 211302 jackiw r , nunez c and pi s -y 2005 _ preprint _ hep - th/0502215 ( to appear in the einstein memorial issue of a ) buchbinder i l , odintsov s d and shapiro
i l 1992 _ effective action in quantum gravity _ ( iop publishing ) birrell n d and davies p c 1982 _ quantum fields in curved space _ ( cambridge university press ) abdalla m c b , nojiri s and odintsov s d 2005 l35 bunch t s and parker l 1979 d * 20 * 2499 parker l and toms d j 1984 d * 29 * 1584 petrov a z 1969 _ einstein space _ , ( pergamon , oxford ) elizalde e , odintsov s d , romeo a , bytsenko a and zerbini s 1994 _ zeta- regularization with applications _ ( singapore : world scientific ) shore g m 1980 376 allen r 1983 228 ishikawa k 1983 d * 28 * 2445 hu b l and oconnor d j 1984 d * 30 * 743 buchbinder
i l and odintsov s d 1985 721 odintsov s d 1993 b * 306 * 233 odintsov s d 1991 _ fortschritte der physik _ * 39 * 621 caldwell r , kamionkowski m and weinberg n 2003 071301 mcinnes b 2002 _ j. high energy phys .
_ jhep0208(2002)029 nojiri s and odintsov s d 2004 b * 595 * 1 nojiri s and odintsov s d 2004 d * 70 * 10352 elizalde e , nojiri s and odintsov s d 2004 d * 70 * 043539 nojiri s , odintsov s d and tsujikawa s 2005 _ preprint _ hep - th/0501025 dolgov a d 1982 in `` _ the very early universe _ '' eds .
gibbobs g , hawking s w , and siklos s t , cambridge university press , p.449 ford l h 1987 _ phys .
rev . _ * d35 * 2339 inagaki t , muta t and odintsov s d 1997 _ prog .
suppl . _ * 127 * 93 ferravis m , francaviglia m and volovich i 1994 1505 allemandi g , borowiec a , francaviglia m and odintsov s d 2005 _ preprint _ gr - qc/0504057 | the functional formulation and one - loop effective action for scalar self - interacting theory non - linearly coupled with some power of the curvature are studied .
after the explicit one - loop renormalization at weak curvature , we investigated numerically the phase structure for such unusual @xmath0 theory .
it is demonstrated the possibility of curvature - induced phase transitions for positive values of the curvature power , while for negative values the radiative symmetry breaking does not take place .
the dynamical mechanism for the explanation of the current smallness of the cosmological constant is presented for several models from the class of theories under consideration . |
this work is the third in a series devoted to non- perturbations of semiclassical s in two dimensions , whose classical bicharacteristic flow is periodic on each energy surface .
the previous works in this series are @xcite , and more recently , in collaboration with s. v ngc , the authors have begun a study of the case when the classical flow of the unperturbed operator is no longer periodic but rather possesses invariant lagrangian tori with a diophantine property
see @xcite for the first work in this direction . in this work
, we continue with the perturbed periodic case . after switching on a perturbation of size @xmath11
, the spectrum will be confined to a band of width @xmath12 , and the more precise distribution of s is very much governed by the flow average of the imaginary part of the leading symbol of the perturbation . in the previous works , we studied the s associated to non - critical values of this flow average or to non - degenerate maxima or minima in a suitable sense ( after restriction to the 2-dimensional of trajectories in an energy surface ) . in this paper
we study the remaining generic case , namely that of s associated with a saddle point .
we will work under the general assumptions of @xcite , @xcite , that we now recall .
let @xmath13 denote @xmath14 or a compact real - analytic of dimension 2 .
we shall let @xmath15 stand for a complexification of @xmath13 , so that @xmath16 in the case when @xmath17 . when @xmath18 , let [ red.01 ] p_=p(x , hd_x,;h ) be the weyl quantization on @xmath14 of a symbol @xmath19 depending smoothly on @xmath20 with values in the space of s of @xmath21 in a tubular of @xmath22 in @xmath23 , with [ red.02 ] p(x,,;h)(1)m((x , ) ) there . here
@xmath24 is assumed to be an order on @xmath22 , in the sense that @xmath25 and for some @xmath26 , @xmath27 , [ red.03 ] m(x)c_0x - y^n_0m(y),x , y^4,x - y:=(1+x - y^2)^12 .
we also assume that [ red.04 ] m1,and [ red.05 ] p(x,,;h)~_j=0^p_j,(x,)h^j , h0 , in the space of such s. we make the ellipticity assumption [ red.06 ] p_0,(x,)m((x,)),(x,)c , for some @xmath28 . when @xmath13 is a compact , we let [ red.07 ] p_=_ma_,(x;h)(hd_x)^,be a on @xmath13 , such that for every choice of local coordinates , centered at some point of @xmath13 , @xmath29 is a smooth of @xmath11 with values in the space of s in a complex of @xmath30 .
we further assume that [ red.07.5 ] a_,(x;h)~_j=0^a_,,j(x)h^j , h0 , in the space of such s. the semi - classical principal symbol in this case is given by [ red.08 ] p_0,(x,)=a_,,0(x)^ , and we make the ellipticity assumption [ red.09 ] p_0,(x,)^m , ( x,)t^*m , c , for some large @xmath28 .
( here we assume that @xmath13 has been equipped with some riemannian metric , so that @xmath31 and @xmath32 are well - defined . ) sometimes , we write @xmath33 for @xmath34 and simply @xmath35 for @xmath36 .
assume [ red.010 ] p_=0 in the case when @xmath13 is compact , we let the underlying hilbert space be @xmath37 for some positive real - analytic density @xmath38 on @xmath13 . under these assumptions , @xmath39 will have discrete spectrum in some fixed of @xmath40 , when @xmath41 are small , and the spectrum in this region will be contained in a band @xmath42 .
assume for simplicity that ( with @xmath43 ) [ red.011 ] p^-1(0)t^*mlet @xmath44 be the hamilton field of @xmath35 .
in this work , we will always assume that for @xmath45 : @xmath46 let @xmath47 , so that [ red.013 ] p_=p+iq+o(^2 m ) , in the case when @xmath18 , and @xmath48 in the compact case .
let [ red.014 ] q=1t(e)_-t(e)/2^t(e)/2qth_p dtp^-1(e)t^*m .
notice that @xmath49 are in involution , @xmath50 . in @xcite
, we saw how to reduce ourselves to the case when [ red.015 ] p_=p+iq+o(^2 ) , near @xmath51
. an easy consequence of this , also remarked upon in @xcite , is that the spectrum of @xmath39 in @xmath52 is confined to @xmath53-\delta , \delta [ + i\epsilon ] \langle \re q\rangle _ { { \rm min},0}-o(1),\langle \re q\rangle _ { { \rm max},0}+o(1)[$ ] , when @xmath54 , where @xmath55 and similarly for @xmath56 .
we shall mainly think about the case when @xmath57 is real - valued but will work under the more general assumption that [ h2 ] qpq , in a region of @xmath58 , where @xmath59 .
let @xmath60 .
assume @xmath61 the connectedness assumption is for convenience only and can easily be removed .
then @xmath62 is a symplectic 2-dimensional near the image @xmath63 of @xmath64 .
we consider @xmath65 as an analytic on @xmath66 .
assume @xmath67 then @xmath63 is an @xmath68-shaped curve , and @xmath57 is an analytic in a of that curve ( which is the level - curve of @xmath65 corresponding to @xmath69 ) . in the following , we may assume that @xmath70 for simplicity .
in section [ sectionred ] we shall construct an @xmath71-dependent @xmath72 which is an @xmath71-perturbation of a real @xmath73 , with such that @xmath74 , @xmath75 and where also the @xmath76 is of @xmath77 . here
@xmath78 is an @xmath68-shaped curve with the self - crossing at @xmath79 and @xmath79 is the saddle point for the @xmath80 with @xmath81 . in the present work it seems quite essential to assume that ( in @xcite this assumption was an optional one that permitted to get improved results . ) after further reductions for the lower order symbols , described in section 2 , we get a reduction of @xmath39 near @xmath82 to an @xmath83 with symbol @xmath84 the @xmath85 is only microlocally defined near @xmath86 but that allows us to define ally its s in a rectangle @xmath53-{1\over c},{1\over c}[+i\epsilon ] { -1\over c},{1\over c}[$ ] and they are of the form where @xmath87 are the s near @xmath88 of @xmath89 in the space @xmath90 defined with floquet conditions along the two loops of @xmath91 as in @xcite . here
@xmath92 with @xmath93 , @xmath94 , and @xmath95 appear as action differences when quantizing @xmath73 , while @xmath96 are maslov indices . for @xmath97 ,
let @xmath98 be the real analytic curve formed by the values of @xmath99 .
that @xmath98 is a curve follows from and we see that @xmath98 is of the form where @xmath100 is in a of @xmath88 . also , let @xmath101 be the critical point of @xmath102 ( with @xmath103 ) , and let @xmath104 be the critical point of the principal symbol of @xmath105 appearing in .
clearly , @xmath106 . put @xmath107 and introduce the exceptional boxes for some fixed large @xmath108 .
[ thint.1 ] we make the assumptions above and especially ( [ red.010 ] ) , ( [ h1 ] ) , ( [ h2 ] ) , ( [ h3 ] ) , ( [ h4 ] ) , ( [ int.3 ] ) , and put @xmath109 , @xmath110 .
assume furthermore that @xmath111 .
for @xmath28 large , the s of @xmath39 in @xmath53-{1\over c},{1\over c}[+i\epsilon ] f_0-{1\over c},f_0+{1\over c_0}[$ ] are of the form ( [ int.5 ] ) , where the following can be said about the @xmath87 : the number of @xmath87 in @xmath112 is @xmath113 if @xmath114 , then @xmath115 , with @xmath26 as in ( [ int.8 ] ) .
there is a bijection @xmath116 between the set of these @xmath87 outside @xmath112 and the union of three sets of points away from @xmath112 : @xmath117 , @xmath118 , @xmath119 such that @xmath120 here @xmath121 is a subset of @xmath122 and @xmath123 , @xmath124 are subsets of @xmath125 ( or vice versa but we only stick to the first option for simplicity ) that can be described by bohr - sommerfeld conditions where in the space of s of @xmath126 , that are smooth near @xmath127 in @xmath128 and in @xmath129 for @xmath130 with a `` @xmath131 '' when @xmath132 and `` @xmath133 '' for @xmath134 .
further , @xmath135 here @xmath136 is a renormalized spectral parameter defined by @xmath137 , with @xmath138 given in propositions ( [ prop1d2 ] ) , ( [ prop1d1 ] ) . finally @xmath139 can be described as actions along suitable cycles in the complexified cotangent space , see section [ sectionev ] .
inside the exceptional boxes the s @xmath87 ( for each fixed @xmath140 ) continue to accumulate to roughly at most 5 curves where three of the curves are the extensions of the curves carrying the @xmath121 , @xmath123 , @xmath124 ( defined by replacing @xmath141 by a continuous real parameter ) and one of the new curves , which exists under certain conditions , can be related to barrier top s in dimension 1 .
there are at most 2 and at least one point ( if we exclude degenerate cases ) where three of the curves terminate and form a `` @xmath142 '' . away from those points
we may have crossings of two of the curves ( like for instance the ones carrying @xmath123 and @xmath124 ) .
away from the `` y '' points and with some margin , the distribution of s can be described by bohr - sommerfeld rules as in the theorem , and near the `` y '' points as well as elsewhere , we can get quite detailed estimates for the distribution of s. indeed , the s can be identified with zeros of quite explicitly given s which in most regions appear as the sum of 4 exponential functions , and for such functions it is possible to study the distribution of zeros quite in detail .
( see davies @xcite for inspiring results in this direction and hager @xcite for quite elaborated results obtained in parallel with the present work . )
the appearance of @xmath142-shaped distributions for non- s in one dimension seems to be quite well known and we refer to shkalikov @xcite , redparth @xcite and further references given there , as well as to the recent works by l. nedelec @xcite and e. servat and a. tovbis @xcite .
the @xmath142shaped eigenvalue distribution is also readily observed numerically see figure 1 on this page , and also figure 2 and figure 3 on pages 8 and 9 , respectively .
unfortunately it turned out to be exceedingly difficult to give a concise and precise description of what happens inside the exceptional boxes in the form of a theorem in less than several pages , so instead we refer the reader to the sections [ sectionsk][sectionev ] where this description can be found .
in section [ sectionbt ] we apply our results to the study of barrier top s for potentials of the form @xmath143 , @xmath144 . in the preceding section [ sectionpara ] ,
we make a remark that permits to improve the domain of validity in the direction of small s. this gives an improvement also in the applications to barrier tops in @xcite and allows us in the present work to treat s @xmath145 with @xmath146 for every fixed @xmath147 , while a direct application of theorem [ thint.1 ] would only give the range @xmath148 .
( in this special situation one can say that the lower bound @xmath149 , can be replaced by @xmath150 for every fixed @xmath27 . )
when starting this project , we underestimated the amount of ingredients needed , and in order to keep the work within a reasonable size , we decided to exclude from the paper the very interesting case when there are more than one saddle point on the same connected component of @xmath151 at real energy 0 .
the most important case here is probably the one with 2 saddle points arising because of an anti - symplectic involution ( typically @xmath152 in the schrdinger case ) . we hope to take up at least the 2 saddle point case in a future work ( having settled essentially all heavy technicalities in the present work ) .
we might then also include the interesting case when breaks down at isolated points , leading to orbifolds .
see colin de verdire v ngc @xcite in the case .
we also ran into a somewhat unexpected difficulty .
indeed , for the one - dimensional s @xmath153 , we have a pseudospectral phenomenon leading to an exponential growth of the resolvent norm in important regions near the spectrum of these s. this makes it very important to keep the errors in the reduction to the @xmath85 in exponentially small , so that the accumulated error in the global resolvent constructions remains controlled . in @xcite we avoided that problem by working in naturally adapted norms where the pseudospectral problems disappeared , but that does not seem equally easy to do here .
this refined reduction is carried out in section [ sectionexp ] using quasi - norms from the theory of analytic s originally due to boutet de monvel kre @xcite , in the simplified variant of @xcite .
the price to pay is the apparent necessity to impose the condition and the upper bound @xmath154 , that should be compared to the bound @xmath155 for every fixed @xmath147 in @xcite or even @xmath156 in @xcite .
* acknowledgment .
* we thank s. fujiie and t. ramond for interesting discussions on the relationship with the complex wkb - method .
those discussion may be important for future generalizations .
we also thank the msri , berkeley , for the possibility to advance on this project during the spring of 2003 .
the first author gratefully acknowledges the partial support provided by the national science foundation under grant dms0304970 .
let @xmath157 be a hypersurface which is transversal to the @xmath158-directions and such that @xmath159 can be identified with @xmath160 .
we can then identify @xmath161 with a curve @xmath162 in @xmath159 .
let @xmath163 , where @xmath164 is the unique increasing analytic with @xmath165 such that the @xmath166-flow is @xmath167 periodic with @xmath167 as its minimal period on @xmath168 .
let @xmath169 be a real- , where @xmath91 is an @xmath68-shaped curve , as in the introduction .
the existence of such a map with a suitable @xmath91 follows from theorem 5 in @xcite ( see also @xcite ) , according to which if @xmath170 and @xmath171 are two compact real - analytic symplectic manifolds of dimension 2 , possibly with boundary , which have the same area and such that there exists an orientation preserving analytic diffeomorphism between them , then @xmath170 can be mapped onto @xmath171 by an analytic canonical transformation .
we extend @xmath172 to a @xmath173 with @xmath174 in the following way : extend @xmath159 to an analytic hypersurface @xmath175 in the full phase space which intersects @xmath176 transversally along @xmath159 .
let @xmath177 be the grad - periodic defined near @xmath168 which solves @xmath178 because of the @xmath167-periodicity of the @xmath166-flow , we see that @xmath177 is well - defined up to an integer multiple of @xmath179 . using that @xmath166 and @xmath180 commute
, we notice that if @xmath181 is a point close to @xmath168 , then we can write @xmath182 where @xmath183 is small and @xmath184 is well - defined modulo a multiple of @xmath167 . put @xmath185
. then @xmath186 has the desired properties . as in section 2 of @xcite
, we introduce the triple @xmath187 of action differences , with @xmath188 corresponding to a closed @xmath158-orbit @xmath189 , and @xmath190 , @xmath191 corresponding to the left and right closed orbits of the @xmath68-shaped set @xmath162 .
let @xmath192 , @xmath193 , where @xmath194 is a suitable maslov index .
let @xmath195 be the space of ly defined s @xmath196 in @xmath197 that are multivalued but @xmath198-floquet periodic as in @xcite , section 6 , or as in @xcite .
let @xmath199 be a ly defined unitary as in the cited works . repeating the argument in the beginning of section 3 of @xcite , we may and will assume from now on that the leading perturbation @xmath200 in ( [ red.013 ] ) has already been averaged along the @xmath158flow so that the leading symbol of @xmath201 becomes the operator @xmath202 has the principal symbol [ red.3 ] _
= g()+iq(,x,)+o(^2 ) . at this stage
, we get a complete analogue to the situation in section 3 in @xcite .
proposition 3.2 there extends , and we get a reduction of @xmath203 to an operator @xmath204 , also acting on @xmath205 , with a complete symbol of @xmath77 , and the principal symbol still of the form now completely of @xmath77 . as in section 5 of @xcite , we use series expansions in the @xmath77-variable and get a reduction of ( [ red.4 ] ) to the family of s : where @xmath206 is a fixed maslov index and it is understood that we only consider such values of @xmath140 for which the first argument ( @xmath207 ) in @xmath85 is small .
\a ) in the general case , without any assumption on the subprincipal symbol of @xmath208 , we write the full symbol of @xmath85 as and consider @xmath209 as an additional small parameter .
\b ) when the subprincipal symbol of @xmath208 vanishes , we have the same fact for @xmath85 by the improved egorov property of @xmath210see section 2 of @xcite .
thus @xmath211 and we can write @xmath212 instead of ( [ red.6 ] ) , we get depending on the small parameters @xmath213 . summing up the discussion
so far , we have [ propred1 ] let @xmath39 be as above satisfying the assumptions ( [ red.02 ] ) , ( [ red.03 ] ) , ( [ red.04 ] ) , ( [ red.05 ] ) , ( [ red.06 ] ) , ( [ red.09 ] ) , ( [ red.010 ] ) , ( [ red.011 ] ) , ( [ h1 ] ) , ( [ h2 ] ) , ( [ h3 ] ) , ( [ h4 ] ) . also assume that @xmath214 for some fixed @xmath215
. then there exist @xmath216 in some fixed of @xmath176 , an elliptic @xmath210 of order 0 , with the associated @xmath186 as above , and an @xmath0- @xmath217 of order @xmath88 with principal symbol @xmath218 , such that the has a symbol @xmath219 of @xmath77 , modulo @xmath220 . here
@xmath221 . in the general case
, we have ( [ red.6 ] ) , provided that @xmath222 , and when the subprincipal symbol of @xmath208 vanishes , we have ( [ red.8 ] ) provided that @xmath223 . in this proposition all symbols and phase s are in fixed @xmath224- domains .
the weight @xmath225 in ( [ red.9 ] ) is used to get a first reduction of the principal symbol to the form ( [ red.2])see also ( [ sk.52 ] ) in section 8 .
since some solution operators to the localized 1-dimensional problems later on will have some exponential growth , the decoupling result of section [ sectionred ] has to be sharpened in the sense that we get some exponential smallness control over the remainders .
first we need to recall some notions about s and their associated quasi - norms .
that was introduced in the pioneering work by l. boutet de monvel and p. kre @xcite , but here we shall use the simplified quasi - norms of @xcite .
if @xmath226 is open , a of order 0 is given by the formal expansion , where @xmath227 are in @xmath228 and satisfy the growth condition , to such an @xmath229 , we associate the formal of infinite order , where @xmath230 let @xmath231 , @xmath232 be an increasing of open subsets with @xmath233 , such that @xmath234 with the distance associated to the @xmath235-norm .
let @xmath236 be the smallest constant such that @xmath237 where @xmath238 is the norm from the space of s on @xmath239 to the same space on @xmath240 .
then @xmath241 is finite for @xmath242 small enough , and conversely , the finiteness of @xmath243 for some fixed @xmath242 implies that @xmath244 is a on @xmath245 .
if @xmath246 are s on @xmath228 and we let @xmath247 and @xmath248 denote the associated @xmath0-s for the classical quantization , then the composition of these two s has the symbol @xmath249 using that the s @xmath250 , @xmath251 compose correspondingly , it was shown very simply in @xcite that implying that the composed symbol is also a .
if we prefer to work with the weyl quantization , then the same result and proof as in @xcite remain valid , provided that we modify the choice of the associated infinite order to @xmath252 so that @xmath253 in the following , we will consider symbols depending on additional parameters ; including @xmath0 ( which is then viewed as an parameter ) and consequently the admissible values of @xmath181 will depend on these parameters . when defining @xmath254 in the case when @xmath0 is among the additional parameters , we have in mind some specific representation of the form ( [ exp.06 ] ) . to have @xmath181 tending to 0 as some power of @xmath0and that is what we will encounter , means roughly that we consider gevrey symbols
[ propexp1 ] let @xmath255 be affine and real and let @xmath256 be an analytic symbol of order 0
. then @xmath257\3_\rho \le 2\rho \vert \nabla \ell \vert _
\infty \3 a\vert _ \rho .\ ] ] here and in the following we shall consider @xmath258 $ ] as an @xmath0- of order 0 .
the symbol of @xmath259 $ ] is equal to @xmath260 where @xmath261 is the hamilton field of @xmath262 .
to @xmath229 we associate the infinite order @xmath263 as in ( [ exp.03 ] ) .
similarly , we have @xmath264\leftrightarrow \sum_{j=1}^\infty h^jb_j(x,\xi , d_x ) = { h\over i}\sum_{k=0}^\infty h^kc_k,\ ] ] with @xmath265 in the sense that @xmath266 acts as a on the s of @xmath267 .
thus , we can also express this as where @xmath268 denotes translation in the complex domain by the vector @xmath269 and @xmath100 may be complex .
assume for simplicity that @xmath270 .
then for @xmath271 , @xmath272 and for @xmath273 in a larger domain , we get @xmath274 for @xmath275 , choose @xmath276 , @xmath277 : @xmath278 in other words , @xmath279 and from the cauchy inequality and ( [ exp.1.5 ] ) , we get @xmath280 here we choose @xmath281 so that @xmath282 minimizes @xmath283 .
we find @xmath284 hence , @xmath285 so @xmath286 and @xmath287 in the following , we allow a finite number of families @xmath288 , @xmath289 , @xmath290 as above . if @xmath291 denotes the @xmath181-quasi - norm , defined with the help of the @xmath266-th , we define here it is understood that if @xmath229 is parameter dependent with @xmath0 among the parameters , then we use the representation ( [ exp.06 ] ) when defining each of the quasi - norms @xmath292 .
notice that we still have ( [ exp.04 ] ) .
[ propexp2 ] let @xmath293 be a of order 0 , defined in a finite union @xmath294 of polydiscs with @xmath295 , and let @xmath296 be the corresponding @xmath0-pseudodifferential operator .
let @xmath297 be a quasi - norm of the form ( [ exp.1.6 ] ) with the corresponding family @xmath298 . then for @xmath181 small enough , we have write @xmath299 , where @xmath300 is a of order 0 .
we notice that @xmath301 ( where @xmath302 is viewed as a symbol of order 0 ) , so on the level , we have hence it only remains to treat the contribution to the commutator from @xmath303 .
we may assume we work in a polydisc centered at ( 0,0 ) with the radii @xmath304 .
then , @xmath305 where @xmath306 now choose the classical quantization for simplicity . on the level , where the sum converges in the space of symbols , since @xmath307 ,
and we can allow some shrinking in @xmath308 and choose @xmath242 small enough . using that @xmath309\3_\rho , \3 [ hd_{x_j},a]\3_\rho \,\le 2\rho \3a\3_\rho $ ] , in view of proposition 3.1 , we see that @xmath310\3_\rho \le
( \alpha _ 1r^{\alpha -e_1}\widetilde{s}^\beta + \alpha _ 2r^{\alpha -e_2}\widetilde{s}^\beta + .. + \alpha _
nr^{\alpha -e_n}\widetilde{s}^\beta + \beta _
1r^\alpha \widetilde{s}^{\beta -e_1}+ .. +\beta _
nr^\alpha \widetilde{s}^{\beta -e_n})2\rho \3 a\3_\rho , \ ] ] which can be written more briefly as @xmath310\3_\rho \le ( \partial _ { r_1}+ .. +\partial _ { r_n}+\partial _ { \widetilde{s}_1}+ .. +\partial _ { \widetilde{s}_n})(r^\alpha \widetilde{s}^\beta ) 2\rho \3 a\3_\rho .\ ] ] hence @xmath3112\rho \3 a\3_\rho .\ ] ] _ remark .
the remains valid for the weyl quantization .
indeed , only ( [ exp.4 ] ) has to be modified , by adding a term @xmath312 to the , where @xmath300 is an symbol of order 0 . _ returning to the considerations of section 2 , let us consider the symbol , defined in @xmath313 , where @xmath314 is open .
here we assume either that @xmath100 is a of order 0 or simply that @xmath100 has an expansion in integral powers of @xmath0 in the space of s. we have already seen in section 2 that after a finite number of conjugations , we may assume that @xmath100 is of @xmath77 modulo @xmath315 .
we also assume that @xmath316 . in the general case , we have and when the subprincipal symbol of @xmath208 vanishes , we have @xmath317 , in the two cases , we shall assume respectively that and for some @xmath147 .
then in both cases , we have @xmath318 .
we shall see how to eliminate the @xmath77-dependence by conjugation with a up to an exponentially decaying error .
the problem can be attacked directly , but it seems that we get better remainder estimates if we first reduce ourselves to the case , when this is possible by means of a functional calculus .
indeed , let @xmath319 be the inverse of the map @xmath164 .
it is easy to see that @xmath320 is well - defined in the sense of formal analytic @xmath0-s or equivalently in the sense of composition of s. ( when @xmath100 is merely assumed to have an expansion in powers of @xmath0 , we consider those @xmath0s as additional parameters . ) we also see that @xmath320 as a symbol has the same properties as @xmath321 above , but now with @xmath164 given by ( [ exp.8.5 ] ) .
it will also be easy to return to the original @xmath321 , for if @xmath322 , then at least formally , @xmath323 , and to say that a symbol is of @xmath77 is equivalent to saying that the corresponding commutes with translations in @xmath77 and this latter property is stable under taking s of the . until further notice @xmath164 will be given by ( [ exp.8.5 ] ) .
using proposition [ propexp2 ] , we get @xmath324&=&h{1\over i}\ { g , a\}(t , x , hd_{t , x};h)+r(a),\\ r(a)&=&\epsilon [ \widetilde{q},a ] , \nonumber\end{aligned}\ ] ] where @xmath325 assuming that the @xmath326-quasi - norm is chosen as in proposition [ propexp2 ] .
consider the map where @xmath327 is supposed to be small .
at least formally , we expand and get the expression for the differential an application of proposition 3.2 shows that the @xmath181 quasi - norm of the last term can be estimated by @xmath328 so , if we assume some fixed upper bound on @xmath327 , we can represent the differential of ( [ exp.19 ] ) as and combining this with ( [ exp.15 ] ) , ( [ exp.16 ] ) , we get the expression for the differential : @xmath329 consider the linear where @xmath251 is a of order 0 and it has the solution clearly ( with a convenient choice of the families @xmath240 ) , choose we look for @xmath250 of the form @xmath330 such that @xmath331 is of @xmath77 .
we shall do the construction by successive approximations in a such a way that uniformly during all the steps , to start with , assume , as we may , that @xmath332 in the sense of ordinary symbols .
choose @xmath333 then , using ( [ exp.23 ] ) , ( [ exp.24 ] ) , @xmath334 where @xmath335 is of @xmath77 and @xmath336 . put @xmath337 then @xmath338 and @xmath339 since @xmath340 the procedure will converge geometrically and we get a formal solution @xmath250 with @xmath341 . by construction @xmath342 for @xmath343
and we have defined @xmath331 by ( [ exp.20 ] ) .
we define @xmath344 as a formal symbol of order 0 , by @xmath345 so that @xmath346 in the space of formal symbols .
similarly , we see that @xmath347 satisfies @xmath348 , @xmath349
. we then verify ( [ exp.19.5 ] ) simply by noticing that @xmath350 notice that from the fact that @xmath351 we infer that locally @xmath352 with @xmath353 .
hence @xmath354 , so @xmath355 so @xmath250 is an symbol with @xmath0 replaced by @xmath356 and can be realized with an uncertainty @xmath357see @xcite .
taking @xmath181 as large as possible respecting ( [ exp.17 ] ) , and recalling that we also assume that @xmath181 is , we get the uncertainty this discussion can also be applied with @xmath250 replaced by @xmath358 .
let @xmath251 be such a realization of @xmath358 .
then we can view @xmath250 as an symbol of order @xmath88 by declaring that @xmath0 is an parameter .
let @xmath359 be a parametrix , so that @xmath360 , if we also denote by @xmath359 a realization . from the construction
, it follows that @xmath361 has a symbol which is @xmath77- up to an error of the size ( [ exp.25 ] ) . with this in mind
, we can state [ propexp3 ] we can construct @xmath250 in proposition [ propred1 ] , such that the symbol @xmath85 there is of @xmath77 up to an error which is @xmath362 . here
we assume ( [ exp.8.3 ] ) in the general case and ( [ exp.8.4 ] ) in the case when the subprincipal symbol of @xmath208 vanishes .
we end this section with a heuristic discussion explaining why we eventually will assume that the subprincipal symbol of @xmath208 is zero . after decoupling by eliminating the @xmath77-dependence as above ,
we get a of 1-dimensional s ( [ red.5 ] ) that we consider at the branching level .
if we first consider the general case without any assumptions on the subprincipal symbol , we can expect to have an estimate on the inverse of these s or on the associated grushin s roughly of the order in order to combine everything , we would like this quantity times ( [ exp.25 ] ) to be @xmath363 .
this is clearly not the case . in the case when the subprincipal symbol of @xmath208 vanishes we expect to improve the estimate on the 1-dimensional resolvents or inverses of grushin problems to roughly this leads to the condition @xmath364 which simplifies to the condition which in addition to ( [ exp.8.4 ] ) gives this is the condition on @xmath1 stated in theorem 1.1 .
in this section and the following one , we study certain model s. much of the material is standard and close for instance to @xcite , @xcite , and @xcite ( see also @xcite for the @xmath365-case ) , but our setup is somewhat different , and we need to recollect some of the basic facts before returning to our @xmath39 . consider or equivalently the @xmath366 from proposition 11 in @xcite we recall that the solutions of ( [ tr.1 ] ) in @xmath367 form a 2-dimensional subspace of @xmath368 . for @xmath369 , we express @xmath273 as @xmath370 , for @xmath371 , we express @xmath273 as @xmath372 , for @xmath373 , we express @xmath374 as @xmath375 , for @xmath376 , we express @xmath377 as @xmath378 . here
@xmath379 is the and we observe that ( [ tr.1 ] ) is equivalent to if @xmath380 , we have two solutions @xmath381 of ( [ tr.1 ] ) , given by where @xmath382 is the heaviside , and the general solution to ( [ tr.1 ] ) becomes a linear combination , we see that @xmath383 is the value of a in the upper half - plane that we also denote by @xmath383 , and for @xmath384 , @xmath385 , we get @xmath386 thus for real @xmath387 , @xmath388 which gives similarly , @xmath389 , with @xmath390 so @xmath391 now let @xmath273 be a solution of ( [ tr.1 ] ) , so that using ( [ tr.5 ] ) and ( [ tr.6 ] ) , we get here we want to express @xmath392 in terms of @xmath393 . from ( [ tr.8 ] ) , we get where we also used the reflection identity , recall that @xmath394 is meromorphic with simple poles at @xmath395 , for @xmath396 , and no other poles .
the reflection identity above can also be written @xmath397 and implies that if @xmath398 , then @xmath399 has to be pole , so @xmath400 for some @xmath396 , @xmath401 , which is impossible since we also know that @xmath402 .
hence @xmath394 has no zeros , and @xmath403 is entire .
the transition matrix in ( [ tr.9 ] ) has determinant 1 and is an entire holomorphic function of @xmath404 .
the relation ( [ tr.9 ] ) remains valid therefore for all @xmath405 .
we next compute the transition matrix analogous to the one in ( [ tr.9 ] ) in the semi - classical case , for solutions of this is of course the same equation as ( [ tr.1 ] ) with @xmath406 .
we now require @xmath407 where [ tr.11.5 ] f_hu()=1e^-ix / hu(x)dx=1(h ) .
a simple computation gives we summarize the discussion above in the following proposition .
[ proptr1 ] let @xmath408 be such that @xmath409 .
if @xmath410 is a solution of ( [ tr.11 ] ) then @xmath411 and there exist @xmath412 , @xmath413 , @xmath414 , @xmath415 such that @xmath416 here @xmath417 is the semiclassical fourier transform defined in ( [ tr.11.5 ] ) and the coefficients @xmath413 , @xmath412 can be expressed in terms of @xmath414 , @xmath418 by ( [ tr.12 ] ) .
the transition matrix which occurs in ( [ tr.12 ] ) is entire holomorphic in @xmath419 and has determinant @xmath420 .
we finish this section by the following observation , which will be useful in section 6 .
the @xmath421 has the principal symbol @xmath422 .
for @xmath423 , @xmath424 , define @xmath425 by @xmath426 , @xmath427 , @xmath428 , @xmath429 .
working in the semiclassical limit , define the microlocal null solutions @xmath430 of @xmath431 , for @xmath432 by @xmath433 near @xmath434 , @xmath435 near @xmath436 , @xmath437 near @xmath438 , @xmath439 near @xmath440 , @xmath441 , @xmath442 . here
@xmath443 is the map
@xmath444 associated to @xmath417 .
then a general null solution of @xmath445 can be written either as @xmath446 or as @xmath447 , where ( [ tr.12 ] ) holds .
in this section , we shall derive asymptotic formulas for the entries of the transition matrix ( [ tr.12 ] ) . we shall use the following version of stirling s formula ( @xcite ) , for every fixed @xmath147 .
we apply this in two cases : _ case a. _ we have @xmath448 and @xmath449 a conic of the negative imaginary axis .
then @xmath450 , @xmath451 avoid a conic of @xmath452 and we can apply ( [ as.1 ] ) , to get : @xmath453\\ & = & { \mbox{\rm exp\,}}\left[-{1\over 2}+i{\mu \over h}-i{\mu \over h}(\ln ( -i{\mu \over h})-{h\over 2i\mu } ) + { \cal o}({h\over \mu } ) \right]\\ & = & { \mbox{\rm exp\,}}\left[-{1\over 2}+i{\mu \over h}-i{\mu \over h}\ln ( -i{\mu \over h})+{1\over 2}+{\cal o}({h\over \mu } ) \right],\end{aligned}\ ] ] so in this case we have here and in what follows @xmath454 always stands for the principal branch of the logarithm .
_ case b. _ we have @xmath448 and @xmath136 avoids some conic of the positive imaginary axis .
then we can apply the earlier results with @xmath136 replaced by @xmath455 and get : if @xmath456 , we combine this with the reflection identity : @xmath457 and the fact that @xmath458 in this region , to conclude that it is straightforward to establish a corresponding estimate in the region @xmath459 , and we can summarize both cases in : _ remark . when @xmath136 is real we have in fact , if we first assume that @xmath460 , we get from ( [ as.2_- ] ) , ( [ as.2_+ ] ) : @xmath461,\ ] ] @xmath462,\ ] ] and using that the second quantity is equal to the complex conjugate of the other , we conclude that ( [ as.4.1 ] ) holds in this case . in the case
@xmath463 , we can use the same argument . in this case
( [ as.2_- ] ) , ( [ as.2_+ ] ) give @xmath464,\ ] ] @xmath465,\ ] ] and we can again conclude that ( [ as.4.1 ] ) holds . _ in any closed sector away from @xmath466 , we can use ( [ tr.12 ] ) and ( [ as.2_- ] ) to get in any closed sector away from @xmath467 , we can use ( [ tr.12 ] ) , the reflection identity and @xmath468 , to get using ( [ as.2_+ ] ) , ( [ tr.12 ] ) , we get for @xmath136 away from a sector around @xmath469 : in a sector @xmath470 , we use the reflection identity @xmath471 to get combining the asymptotic formulae ( [ as.5])([as.8 ] ) , we may summarize the discussion in this section in the following proposition .
we have the following tableau for the coefficients of the transition matrix ( [ tr.12 ] ) , when @xmath472 . in all cases :
@xmath473 for @xmath474 : @xmath475 @xmath476 for @xmath477 : @xmath478 @xmath479 for @xmath480 : @xmath481 @xmath482 for @xmath483 : @xmath484 @xmath485 here the terms with @xmath486 in the exponents indicate that we should take the sum of the two possible terms with the same remainders @xmath487 or @xmath488 in the exponent for each of the two terms .
we now return to proposition [ propred1 ] , which shows ( when combined with the exponential remainder estimates of section [ sectionexp ] ) , that the study of @xmath39 near @xmath489 ( considered in a suitable weighted space ) can be reduced by conjugation to that of @xmath85 acting on the space @xmath490 of s defined ly in some fixed of @xmath491 , @xmath492 in @xmath493 , with a @xmath494 floquet periodicity condition . here
@xmath91 is an @xmath68-shaped curve as in the introduction , and @xmath495 was defined in the beginning of section [ sectionred ] . in order to fix the ideas
, we assume that we are in the case when the subprincipal symbol of @xmath208 vanishes , so that the symbol of @xmath85 is given by ( [ red.8 ] ) , and we are then in the parameter range @xmath496 .
at least formally , the study of @xmath85 can be reduced to a of one - dimensional s by a series expansion in the @xmath77-variable and ( as noted in ( [ red.5 ] ) ) we get the one - dimensional s where @xmath497 , and we restrict the range of @xmath140 by requiring that @xmath498 be small .
the s ( [ 1d.1 ] ) should be considered as acting on the space @xmath90 defined similarly to @xmath195 , with @xmath499 . using ( [ red.8 ] ) , we see that ( [ 1d.1 ] ) becomes where @xmath500 is @xmath501 in a fixed complex of @xmath502 and depends smoothly on the other parameters .
we further notice that is equal to the average of @xmath200 , expressed in suitable real symplectic coordinates , and we know by construction that we also recall the assumptions ( [ h2 ] ) , ( [ h4 ] ) , which say that here we assume for simplicity that @xmath503 is the branching point .
in the following we shall discuss the spectrum of the 1-dimensional since this is only defined ly and up to an error @xmath504 , we have to keep in mind that for the moment the s will be defined only formally and up to errors of at least the same size . _ a first localization of the spectrum
. assume first that @xmath57 is real - valued .
then from the sharp grding inequality , we see that the spectrum of the ( [ 1d.7 ] ) in the band @xmath505 is contained in under the more general assumption ( [ 1d.5 ] ) , we see that ( [ 1d.8 ] ) can be applied to @xmath506 , where @xmath507 .
so in the general case , we see that the spectrum of the ( [ 1d.7 ] ) in the band @xmath505 is contained in we also have _ _ normal forms near the branching points .
_ let @xmath508 be the principal symbol of @xmath153 .
following @xcite , @xcite , we get the following adaptation of proposition 5.3 of @xcite : [ prop1d1 ] we can find a : @xmath509 depending ally on @xmath207 and smoothly on @xmath213 with values in the s : @xmath510 , and an @xmath511 depending smoothly on @xmath512 , such that @xmath513 is the unique critical point close to ( 0,0 ) of @xmath514 and with moreover , @xmath515 is real when @xmath207 is real and @xmath516 after a conjugation by an elliptic associated to the @xmath517 we may assume that the leading symbol of @xmath518 , is a of @xmath519 , @xmath1 , @xmath520 and @xmath521 .
we can get a complete normal form by making further conjugations by analytic s of order 0 in such a way that the complete symbol also becomes a of @xmath519 , @xmath1 , @xmath520 and @xmath521 .
this is carried out in appendix a. we get the following result which is very close to one of the main results of the appendix b in @xcite .
[ prop1d2 ] we can quantize @xmath517 by an elliptic @xmath522 with an analytic symbol , depending ally on @xmath207 and smoothly on @xmath213 , such that where @xmath523 is a classical analytic symbol of order @xmath88 depending ally on @xmath207 and smoothly on @xmath213 . the leading part of @xmath138 is @xmath524 appearing in proposition [ prop1d1 ] . _
the quantization condition .
we start with a side remark about normalization .
when @xmath525 is a one - dimensional @xmath0- of principal type , and @xmath526 , then we can normalize ly defined solutions of @xmath527 , @xmath528 , by imposing that @xmath529u\vert u)=1.\ ] ] here @xmath530 is defined near a piece of the real characteristics and has the property that @xmath531 is of compact support near the characteristics of @xmath532 and @xmath533 increases from 0 to 1 when we progress in the hamilton flow direction .
( see @xcite . )
it is easy to check that if we view @xmath273 as a solution of @xmath534 , then we get the corresponding normalization @xmath535u\vert u)=g(z , z)({i\over h}[p,\chi ] u\vert u)=g(z , z),\ ] ] where @xmath536 if we drop the requirement that @xmath321 be or just let @xmath537 become complex , there is no obvious normalization of null - solutions of @xmath532 , but we still have a well - defined sesqui - linear form on @xmath538 , given by @xmath529u\vert v).\ ] ] if we have some additional information allowing us to identify the two null - spaces , then this can still be used to normalize null - solutions of @xmath532 . in the following
we abandon the attempt to normalize completely the null - solutions , since already the @xmath539 is not necessarily . _ by proposition [ prop1d2 ] we have an analytic symbol @xmath540 depending analytically on @xmath207 and smoothly on @xmath71 , @xmath520 , such that notice that if @xmath273 is a null - solution of @xmath445 in a full of @xmath79 , then @xmath541 near @xmath79 , where the spectral parameters are related by recall from the end of section [ sectiontr ] that @xmath445 has the four characteristic points @xmath542 , @xmath543 and that this has the null solutions @xmath430 described after proposition [ proptr1 ] .
when @xmath136 is real , we check that @xmath430 is normalized near @xmath542
. if @xmath544 is a global null - solution of @xmath445 , with @xmath545 near @xmath542 , then by proposition 4.1 we have : assume for simplicity that @xmath546 is defined in a suitable domain , containing @xmath542 , @xmath432 .
let @xmath547 , @xmath548 .
then if @xmath273 is null - solution of @xmath549 near the branching point @xmath550 , equal to @xmath551 near @xmath552 , then ( [ 1d.35 ] ) still holds . we may wish to renormalize the @xmath553 , by putting
then a straightforward calculation shows that if @xmath273 is a null - solution of @xmath549 near the branching point , and @xmath554 near @xmath552 , then here is a natural example of such a renormalization .
assume for simplicity that near @xmath552 the set @xmath555 takes the form @xmath556 , where @xmath557 is and depends ally on the parameters @xmath558 .
then choose @xmath559 so that ly near @xmath552 we have the standard wkb - form : where @xmath560 is a classical elliptic analytic symbol of order 0 .
the function @xmath561 solves the eikonal @xmath562 and @xmath563 depend ally on the additional parameters @xmath564 and smoothly on @xmath565 .
using an explicit representation of @xmath210 we write near @xmath552 for @xmath566 : here the last equality follows from stationary phase , @xmath567 is a phase generating @xmath186 and near @xmath542 we write @xmath430 in the wkb - form @xmath568 the function @xmath569 in ( [ 1d.36 ] ) appears as the critical value in the stationary phase expansion of ( [ 1d.36 ] ) and solves the same eikonal equation as @xmath561 . for @xmath570
, we get near @xmath552 : where @xmath571 and @xmath572 appear when writing @xmath573 on wkb - form near @xmath574 . in this case
we see that @xmath575 is a classical analytic symbol of order 0 , depending ally on the additional parameters and with the imaginary part of the leading symbol vanishing when @xmath576 , @xmath577 , @xmath578 .
the leading part of @xmath579 can be further described in terms of symplectic geometry .
put we have the obvious relation now we work in a full of @xmath91 ( introduced after ( [ int.2 ] ) ) . recall that we have the points @xmath580 , @xmath581 , @xmath582 , @xmath583 on the four crossing branches of @xmath91 distributed with positive orientation around the branching point .
we may assume that @xmath582 , @xmath583 are situated close to the left closed curve @xmath584 of @xmath91 and that @xmath580 , @xmath581 are situated close to the right closed curve @xmath585 of @xmath91 .
start with a null - solution to @xmath549 near @xmath550 , of the form @xmath586 near @xmath587 and of the form @xmath588 near @xmath589 . here for the moment @xmath393 can be prescribed , and we then know that @xmath554 near @xmath552 for @xmath590 , where @xmath413 , @xmath412 are given by ( [ 1d.41 ] ) .
we require temporarily that @xmath273 is a well - defined single - valued null - solution along the whole left closed component @xmath591 of @xmath91 .
then if we follow @xmath273 around the exterior part of @xmath584 from @xmath587 to @xmath589 , we get with @xmath592 denoting the exterior part of @xmath591 which joins @xmath587 to @xmath589 .
now recall that we really want @xmath593 , @xmath594 , where @xmath595 where @xmath596 is a real action difference related to the reduction in proposition [ propred1 ] and @xmath94 a corresponding maslov index .
this means that @xmath273 should be multivalued , but floquet periodic along @xmath591 in the sense that where @xmath597 denotes the extension of @xmath273 along one loop of @xmath591 which we assume to be oriented in the following way : @xmath598 .
starting near @xmath589 , we get @xmath597 near the same point in two steps : @xmath599 the floquet condition ( [ 1d.45 ] ) therefore becomes @xmath600 , or equivalently instead of ( [ 1d.44 ] ) .
similarly , let @xmath601 be the right hand loop in @xmath91 with the orientation : @xmath602 . then
, if we want @xmath273 to extend to a null - solution in @xmath603 near @xmath601 , we get the analogue of ( [ 1d.46 ] ) : with @xmath604 defined as in ( [ 1d.44 ] ) with @xmath592 there replaced by @xmath605 , the exterior segment in @xmath606 that joins @xmath607 to @xmath608 .
start near @xmath587 with @xmath586 , use ( [ 1d.46 ] ) to get @xmath414 and then ( [ 1d.41 ] ) to get @xmath392 : and in order to get a global solution in @xmath609 , we also need to apply ( [ 1d.47 ] ) , which gives our global one - dimensional quantization condition where we took @xmath610 . in this relation ,
we substitute ( [ 1d.41 ] ) , ( [ 1d.42 ] ) and get after multiplication with @xmath611 : @xmath612 with @xmath613 where we recall that @xmath614 with [ 1d.49.2 ] _ j , k=_j , k+h2 , we get [ prop1d3 ] assume that @xmath615 is holomorphic in @xmath616 and depends smoothly on @xmath617 . here
@xmath91 is an @xmath618shaped curve with the self - crossing at @xmath79 .
assume furthermore that @xmath619 where @xmath620 is an analytic function with @xmath621 .
we assume next that along @xmath91 , @xmath622 and that @xmath623 has a unique critical point on @xmath91 , @xmath79 , which is a non - degenerate saddle point .
when @xmath624 , put @xmath625 where @xmath626 is an analytic symbol introduced in propositions 6.1 and 6.2 , and in ( [ 1d.34 ] ) .
then @xmath537 is a quasi - eigenvalue of the operator @xmath627 acting on @xmath628 if and only if the corresponding @xmath419 satisfies ( [ 1d.50 ] ) . in ( [ 1d.50 ] ) , the coefficients @xmath629 , @xmath630 , @xmath631 , and @xmath632 are introduced in proposition 4.1 , and the quantities @xmath633 and @xmath634 are defined in ( [ 1d.49.1 ] ) and ( [ 1d.49.2 ] ) .
they depend ally on @xmath419 with and when @xmath136 is real , we have in the formulation of the proposition , we leave the notion of a quasi - eigenvalue undefined and refer the reader to section 11 for a complete justification of this terminology .
here we elaborate on arguments in @xcite , and a related and even more general discussion can be found in hager @xcite .
the results established in this section will be used in section 10 .
let @xmath635 be compact @xmath636 segments in @xmath637 such that @xmath638 starts at @xmath639 and ends at @xmath640 , where we use the cyclic convention and view the index @xmath641 as an element of @xmath642 .
we assume that @xmath643 is fixed , but allow @xmath644 to vary with the semi - classical parameter @xmath0 while all estimates below will be uniform in @xmath0 .
let @xmath620 be a defined in @xmath645 , such that @xmath646 @xmath647 where @xmath596 is in @xmath648 and @xmath649 in @xmath650 we can write @xmath651 we further know that @xmath652 . standard arguments ( see for instance @xcite ) , including jensen s formula , imply that the number of zeros of @xmath559 in @xmath653 is @xmath654 and if @xmath552 is a segment in @xmath653 from @xmath430 to @xmath655 which avoids the zeros @xmath656 of @xmath559 in @xmath653 such that @xmath657 for every @xmath140 , then and consequently let @xmath658 be the closed contour given by @xmath659 .
we want to study @xmath660 when @xmath658 is the oriented of a domain @xmath661 , where @xmath620 is , then @xmath662 is the number of zeros of @xmath620 inside @xmath661 .
along @xmath638 , we write @xmath663 , @xmath664 . then , @xmath665 so @xmath666 and hence in view of ( [ ze.2 ] ) and the uniform ness of @xmath667 : here we recall that @xmath668 in @xmath650 .
it follows that @xmath669 in @xmath653 and consequently that @xmath670 for any other point @xmath671 . thus finally , with @xmath672 chosen .
here we can further replace @xmath673 by its real part , since @xmath674 in @xmath650 .
we now return to the situation in section [ section1d ] .
we are interested in the solutions @xmath136 of ( [ 1d.50 ] ) . in the following ,
we will write @xmath676 instead of @xmath677 , so we are interested in the zeros of the @xmath678 appearing in ( [ 1d.50 ] ) , given by @xmath679 pulling out a factor @xmath680 , we get the new equivalent function which has the same zeros as @xmath69 . here
we have also used the explicit formulae for @xmath632 , @xmath629 in ( [ tr.12 ] ) . using the results of section 5
, we shall now look at the s of @xmath681 , when @xmath682 .
_ case 1 : _ assume that ( the case 2 , given by @xmath683 will be reduced to the case 1 by a symmetry argument . ) in this region , we have ( [ as.8 ] ) : @xmath684 and we get using also ( [ sk.1 ] ) and ( [ as.5 ] ) , @xmath685 where @xmath686 we have where @xmath687 notice that @xmath688 is dominating over @xmath689 when @xmath690 , and in each half - plane @xmath691 , we may associate @xmath692 to the dominating term , modulo an error which is @xmath693 times the leading term .
also notice that the last equations take the form @xmath694 where following the general principles , as explained for example in @xcite ( see also @xcite ) , we shall now look for the curves @xmath695 , @xmath696 , where @xmath697 , and we shall especially be interested in those parts of @xmath695 where @xmath697 is dominating over the other @xmath698 . in doing so ,
let us remark first that we will not see any zeros of @xmath699 generated by the zeros of @xmath700 as a dominating part of @xmath699 , for when @xmath701 , then @xmath702 dominates over @xmath703 and @xmath704 and clearly we can not have @xmath705 .
now when @xmath706 , @xmath703 dominates over @xmath702 and @xmath707 so that @xmath708 in this case , leading to an even stronger conclusion .
we now begin look at the location of zeros of @xmath692 and of sums of two of the @xmath571 .
zeros of @xmath692 : they are of the form @xmath709 , @xmath710 . zeros of @xmath711 : they are contained in the region @xmath712 : similarly the zeros of @xmath713 are contained in @xmath714 : the zeros of @xmath715 are contained in @xmath716 : the zeros of @xmath717 are contained in @xmath718 : the zeros of @xmath719 are contained in @xmath720 : put @xmath721 when @xmath722 , @xmath723 dominates over @xmath724 and we shall only consider @xmath725 , @xmath726 , @xmath727 given by recall now from section 5 that @xmath728 in ( [ sk.3.1 ] ) appears as a remainder in stirling s formula , so that @xmath729 , and hence @xmath730 is lipschitz for @xmath731 .
proposition [ propa3.1 ] can therefore be applied to get the approximate behavior of the @xmath732 .
this will be exploited later . in the left half - plane
, @xmath724 dominates over @xmath723 and we consider all the 5 curves @xmath733 , @xmath734 , @xmath735 , @xmath718 , and @xmath716 , given by : again proposition [ propa3.1 ] can be applied to give the approximate shape of @xmath732 . recall that we are in the case 1 with @xmath675 , so that ( [ sk.6 ] ) holds .
_ 1 . skeleton in the region @xmath736 .
( we will implicitly use that @xmath737 . ) the region @xmath738 will require a special discussion . in the region @xmath739 , we have @xmath740 , so @xmath741 and in this region we see from an earlier observation that the zeros of @xmath700 will not play any essential role . in this region
we shall therefore use the @xmath732 appearing in ( [ sk.4 ] ) and , as pointed out earlier , we are interested here in the part of each such @xmath732 , where @xmath742 dominates over the other @xmath743 . _
it follows from proposition [ propa3.1 ] that the curves in ( [ sk.4 ] ) ( as well as the ones in ( [ sk.5 ] ) ) , are of the form notice that every crossing point of two of the curves @xmath744 is also a crossing point for all three .
this follows from ( [ sk.4 ] ) or even more trivially from the observation that 2 of the 3 s @xmath745 , @xmath746 , @xmath747 , imply the third one .
also notice that if we draw the two curves @xmath748 , @xmath749 , then @xmath727 is between the two see figure 2 on page 35 .
for @xmath750 we have @xmath751 by ( [ as.4.1 ] ) and hence @xmath752 is described as in proposition [ propa3.1 ] with in the region @xmath753 we have @xmath754 , and if we restrict further to with @xmath4 , we see that @xmath723 is dominating in the sense that and hence @xmath755 has no zeros in that region . similarly in the region @xmath756 , we have @xmath757 , and if with @xmath4 , then @xmath758 is dominating in the sense that and again @xmath755 has no zeros there .
now consider a point @xmath759 , where @xmath760 , so that @xmath761 , with @xmath762 . going down ( ie .
decreasing @xmath763 while keeping @xmath764 constant ) by a distance @xmath765 , we reach the region ( [ sk.9 ] ) , where @xmath723 dominates . _
case a. @xmath766 .
going up by a distance @xmath767 , we cross @xmath768 and reach the region ( [ sk.11 ] ) , where @xmath758 is dominating . _ _ case b. @xmath769 for @xmath770 . going up by a distance @xmath771
, we reach the region , where @xmath772 is dominating : and continuing to go up , @xmath772 remains dominating until we reach a @xmath773- of @xmath749 . after crossing that curve and going up by another amount @xmath774 , we reach the region , where @xmath758 is dominating . _ [ l][l]@xmath775 [ l][l]@xmath776 [ 1][1]@xmath777 [ 1][1]@xmath778 [ 1][1]@xmath779
[ 1][1]@xmath780 [ 1][1]@xmath723 dominates [ 1][1]@xmath758 dominates [ 1][1]@xmath781 dominates [ 1][1]@xmath772 dominates our discussion shows the following [ propsk.1 ] we work in the region ( [ sk.6 ] ) and assume in addition that @xmath782 . if @xmath783 , then if @xmath784 , then if @xmath785 , then if @xmath786 , then if the distance from @xmath136 to @xmath787 is @xmath788 , with @xmath4 , then in the respective cases ( [ sk.15])([sk.17 ] ) can be sharpened to the dominance in the sense explained above . in particular , @xmath699 has no zeros in this region . if , in addition , @xmath789 with @xmath4 , then we have the same conclusion in the case of ( [ sk.14 ] ) .
in the region ( [ sk.6 ] ) , intersected with the right half - plane @xmath722 we define the skeleton to be the union of the curves @xmath790 and @xmath791 .
the proposition shows that the zeros of @xmath699 in the region under consideration are contained in the union of all discs @xmath792 with @xmath136 in the skeleton just defined , and the set of all @xmath419 below the skeleton , with @xmath793 , for @xmath4 .
skeleton in the region @xmath794 .
again the region @xmath795 will require a separate discussion so we restrict the attention to @xmath796 and we will use @xmath797 , so that @xmath798 .
we therefore concentrate the attention to the curves in ( [ sk.5 ] ) . as before we notice that every crossing point of two of the three curves @xmath799 is a crossing point of all three .
the same holds for @xmath800 .
_ now use that @xmath801 is small and hence that @xmath802 and their derivatives @xmath764 are small
. we can therefore consider the two curves : @xmath803 they are of the form @xmath804 where @xmath805 , @xmath806 are small with small derivatives and satisfy : the curve @xmath807 will play a central role .
it crosses @xmath808 at the points @xmath809 ( unless these points are hidden in the forbidden region ) , and we have we notice that @xmath810 is the unique crossing point for @xmath811 while @xmath812 is the unique crossing point for @xmath813 .
more precisely , @xmath814 , @xmath815 vanish precisely for @xmath816 and have the same sign as @xmath817 .
similarly @xmath818 , @xmath819 vanish precisely for @xmath820 and have the same sign as @xmath821 .
we also notice that if @xmath419 belongs to one of the three curves @xmath822 , then the distance from @xmath136 to any of the two other curves among these three is @xmath823 .
the same observation holds for @xmath824 with @xmath812 instead of @xmath810 . for @xmath825
, we have @xmath826 by ( [ as.4.1 ] ) , and hence @xmath827 is described as in proposition [ propa3.1 ] with now assume to fix the ideas that @xmath828 ( the case @xmath829 can be treated similarly ) .
considering the three curves @xmath822 , we see that for @xmath794 : if @xmath830 , then @xmath831 ( in this case , we also have @xmath832 . ) if @xmath833 , then @xmath834 ( in this case , we also have @xmath835 . ) if @xmath836 , then @xmath837 this covers all possible cases with @xmath794 .
notice that the last case can appear only when @xmath838 .
when @xmath829 , we get the analogous discussion after a permutation of the indices 2 and 3 : if @xmath839 , then @xmath831 if @xmath840 , then @xmath834 if @xmath841 , then @xmath842 moreover , if in addition all the inequalities for @xmath843 are valid with an extra margin @xmath774 , @xmath4 , then in the various cases , we have dominance of @xmath844 respectively , in the sense explained before . for the dominance of @xmath724 , we also need the assumption that @xmath796 with @xmath4 .
[ l][l]@xmath775 [ l][l]@xmath776 [ 1][1]@xmath845 [ 1][1]@xmath846 [ 1][1]@xmath847 [ 1][1]@xmath848 [ 1][1]@xmath849 [ 1][1]@xmath850 [ 1][1]@xmath851 dominates [ 1][1]@xmath724 dominates [ 1][1]@xmath772 dominates _ exponential localization to the skeleton .
recall that we are still working in the region ( [ sk.6 ] ) . in this region
we define the skeleton to be : where we define @xmath852 to be the union of the following two sets in respectively the left and the right half - planes : _ in the closed left half - plane ( intersected with ( [ sk.6 ] ) ) , assume @xmath828 to fix the ideas , then this part of @xmath852 is given by all points of the form @xmath853 with @xmath854 , all points of the form @xmath855 or of the form @xmath856 , with @xmath838 .
in the closed right half - plane the corresponding part of @xmath852 is defined to be the union of the two curves : @xmath857 and @xmath858 .
@xmath859 is defined to be the part of the imaginary axis given by @xmath860 where @xmath861 is the same constant as in ( [ sk.6 ] ) .
notice that this part may be empty .
the earlier discussion shows that we have [ propsk.2 ] the zeros of @xmath699 in the domain ( [ sk.6 ] ) are contained in the set .
the localization result of proposition 8.2 improves if we are far from the branch points of the skeleton .
thus for instance , if @xmath136 is below @xmath852 and @xmath862 , then
@xmath863 so the zeros of @xmath699 are exponentially small perturbations of those of @xmath864 : in this region there is a bijection @xmath865 between the zeros of @xmath692 and those of @xmath699 , with @xmath866.\ ] ] similarly let @xmath867 be a point of @xmath852 in the right half - plane , say @xmath868 with @xmath869 dominating above and below this point respectively . if @xmath870 then in @xmath871 , we have @xmath872 and we conclude that the zeros of @xmath699 are exponentially close to those of @xmath715 .
essentially the same statement holds when @xmath867 belongs to the lower part of @xmath852 , but here the size of @xmath764 also matters , so near a point @xmath873 we get a bijection @xmath865 between the zeros of @xmath874 and those of @xmath699 with @xmath875 and we have to assume both that @xmath876 and that @xmath877 .
the analogous statements hold in the left half - plane . _ more refined analysis in the region @xmath878 .
the study of the upper part ( @xmath716 or @xmath718 ) of the skeleton is unchanged in this region , while the lower part requires more attention in view of the fact that @xmath879 may be considerably smaller than @xmath880 when we are close to a zero of @xmath692 . in order to fix the ideas we assume that @xmath881 . _ after multiplication of @xmath882 by the same exponential factor , we arrive at and we shall drop the tildes in the following discussion . here with we have @xmath883 and recall that we work in the region @xmath884 . notice that @xmath885 .
similarly , we look at for a suitable branch of the square root .
also @xmath886 , so this relation gives a bound for @xmath887 and its real part ; @xmath888 .
we have the general estimate so we get assuming we conclude that and consequently , where @xmath654 denotes terms that are and @xmath889 denotes terms that tend to 0 , when @xmath890 for each zero @xmath891 of @xmath892 , we introduce the diamond shaped with @xmath861 large enough so that the preceding estimates apply away from the union of all the @xmath893 . define @xmath894 to be the set of points with @xmath895 away from the union of all the @xmath893 , with @xmath896 added , if @xmath896 has the property that the distance from this diamond to the points just defined , is zero .
@xmath896 is unique if it exists , since the other points of @xmath894 form a curve @xmath897 , with @xmath898 . from the above estimates we get ( if @xmath894 stays away from an @xmath899- of the zeros of @xmath892 , then the agreement is better : @xmath900 in fact
, we can get an even more precise estimate for the distance between @xmath894 and @xmath712 : let @xmath901 , and put @xmath902 then @xmath903 away from the diamonds .
we therefore get the following estimate for the vertical distance from @xmath867 to @xmath712 : ( assuming for simplicity that @xmath904 ) .
this is a refinement of the lower bound in ( [ sk.32 ] ) , and the argument also gives the upper bound there .
we reach the following conclusion about the location of the zeros in the region @xmath905 : [ propsk.3 ] above @xmath852 and at distance @xmath906 from @xmath852 , @xmath907 is dominating .
@xmath692 is dominating if @xmath136 is below @xmath852 , at distance @xmath906 from @xmath908 and at distance @xmath909 from @xmath852 . in between ( for instance below @xmath716 but above @xmath712 ) , @xmath772 ( or @xmath781 ) is dominating if the distance to the skeleton is @xmath910 .
_ improvement in the region @xmath762 .
let us recall that therefore , the zeros of @xmath911 are situated on @xmath912 and are given by the explicit quantization condition the distance between successive zeros is @xmath771 .
if @xmath867 is such a zero , then in a disc @xmath913 with @xmath914 , we have @xmath915 away from the union of all such discs , we have @xmath916 similarly , the zeros of @xmath917 are situated on the curve @xmath918 and given by the quantization condition and the other statements about @xmath911 carry over to @xmath917 .
_ now consider we get [ propsk.4 ] in the region @xmath762 , there is a bijection @xmath865 from the union of the zeros of @xmath919 and of @xmath920 to the zeros of @xmath699 with so in the region @xmath762 , and modulo an exponentially small error , we can identify the zeros of @xmath699 with the union of the zeros of @xmath919 and of @xmath920 .
this finishes the analysis of the skeleton in the first case ( [ sk.6 ] ) .
_ case 2 : _ assume that in this case from ( [ as.6 ] ) and ( [ as.7 ] ) , respectively , we recall that @xmath921,\ ] ] and @xmath922.\ ] ] using ( [ sk.1 ] ) , we get where @xmath571 , @xmath543 are the same as in case 1 , but now with a partially different representation : @xmath923 again we consider @xmath0 times the real parts of the different exponents of the @xmath571 , @xmath924 : @xmath925 where by a symmetry argument , we shall now see that @xmath926 plays the same role in the present case 2 as @xmath927 in case 1 , provided that we perform the following transformations on @xmath926 : \1 ) add @xmath928 to each of the five terms .
\2 ) replace @xmath136 by @xmath929 .
then we get @xmath930 : @xmath931 this is analogous with @xmath927 in case 1 except for the fact that @xmath932 here corresponds to @xmath933 in case 1 .
_ remark . using ( [ as.4 ] ) , it is easy to check that when it follows that when @xmath136 satisfies ( [ sk.43 ] ) , and hence if @xmath136 belongs to the skeleton @xmath852 defined according to case 1 , the distance from @xmath136 to the corresponding skeleton @xmath852 defined according to case 2 is @xmath934 _
we end this section by some general considerations that will be useful in section [ sectionbt ] .
we see from ( [ sk.8.5 ] ) that the spectrum will have a genuinely 2-dimensional structure if or if @xmath935 and @xmath936 have the same sign and in the latter case , we even have some s on the imaginary @xmath136-axis , related to 1-dimensional barrier top s. it is therefore important to have a invariant and direct description of @xmath937 , @xmath936 .
the final definition of @xmath938 in the beginning of section 6 is simply that we start with the null - solution @xmath939 of @xmath153 near @xmath587 and extend it along the exterior part of @xmath91 until we reach a of @xmath582 , where we get @xmath940 .
( here we neglected the real floquet parameter @xmath941 , since we are only interested in the imaginary part of @xmath938 ) .
the definition of @xmath604 is similar .
now take @xmath942 ( cf .
( [ 1d.33 ] ) and ( [ 1d.34 ] ) ) and represent the @xmath153 as acting in @xmath943 , where @xmath944 is , with @xmath945 a of @xmath91 in @xmath946 . from the construction of @xmath430 , @xmath553
, we see that @xmath553 is near @xmath552 a normalized null - solution of @xmath153 in @xmath947 , where @xmath948 is small and @xmath949 is defined in a large of the projection of the branching point . here
if @xmath950 is the canonical transformation associated to some standard fbi - bargmann transform .
then @xmath951 , @xmath952 , with @xmath210 as in ( [ 1d.33 ] ) .
since @xmath953 is naturally identified with @xmath946 , where @xmath954 , so ( since @xmath942 ) , we know that the null set of @xmath955 intersects @xmath956 along two crossing `` real '' curves that we we can identify with `` the interior part '' of @xmath91 .
extend @xmath949 to be defined in @xmath228 with @xmath948 small . if @xmath957 , then where @xmath592 now ( cf ( [ 1d.44 ] ) ) is a real curve from @xmath587 to @xmath589 in @xmath958 close to the exterior part of the `` left loop '' of @xmath91 .
here we let @xmath959 denote the @xmath537-value in corresponding to @xmath960 .
let this left loop be denoted by @xmath591 and let us consider it ( after slight deformation ) as a closed curve in @xmath958 joining the critical point of @xmath955 to itself staying close to the left loop of @xmath91 . here
, we may assume that the interior part of @xmath584 ( joining @xmath589 to @xmath587 ) is contained in @xmath956 , so @xmath961 there .
hence ( [ sk.47 ] ) becomes here @xmath962 in the last integral can be replaced by @xmath963 , which by stokes formula can be further replaced by any other real 1-form @xmath964 with @xmath965 , @xmath966 .
this means that we can reinterpret the last integral in ( [ sk.48 ] ) as the corresponding one along the corresponding closed curve in the complexification of @xmath14 . to simplify things further , recall that @xmath11 , @xmath967 are small perturbative parameters for @xmath955 and that @xmath958 is real when @xmath968 . in general , if @xmath969 depends smoothly on a real parameter @xmath970 , if @xmath145 is not a critical value and @xmath971 is a simple closed curve in @xmath972 , then for @xmath145 fixed ; where @xmath973\ni t\mapsto { \mbox{\rm exp\,}}th_{q_s}(\rho ( 0))$ ] is a natural parametrization .
this can be applied to the case @xmath974 , so if @xmath145 also depends on @xmath970 , we get @xmath975 in this form , we can treat a loop like @xmath976 starting and ending at the critical points @xmath977 of @xmath978 , parametrized by @xmath53-\infty , + \infty [ \mapsto { \mbox{\rm exp\,}}(th_{q_s})(\rho ( 0))$ ] , provided that we take @xmath979 equal to the critical value @xmath980 : this can be proved by a limiting procedure , approaching @xmath981 by closed curves at non - critical levels .
taking the imaginary parts , this means that we have a fairly simple way of computing @xmath982 , @xmath983 perturbatively . from this computation
, we see that it is of interest to compute the @xmath984 contribution to the averaged principal symbol .
this computation was carried out in section 2 of @xcite under the assumption that @xmath985 and it works essentially the same way without that assumption : we start with the principal symbol the function introduced in proposition [ propred1 ] , satisfies put @xmath986 , where @xmath987 remains to be determined . as in @xcite we get at a general point @xmath988 , ( @xmath989 ) : @xmath990 where we used that @xmath991 . letting @xmath987 solve @xmath992 we get with @xmath993 ,
now assume for simplicity that @xmath994 is constant .
then @xmath995 here , we notice that @xmath996 so finally : the formulas ( [ sk.54 ] ) , ( [ sk.55 ] ) will be used in section [ sectionbt ] together with the following remark : if we put then a simple computation shows that if we put then combining ( [ sk.57 ] ) and the @xmath997 periodicity of @xmath998 with the change of variables @xmath999 , we get
in this section we shall consider the case @xmath1001 . in doing so , we will use ( [ tr.12 ] ) more directly .
we will first work in a region @xmath1002 , where @xmath1003 , @xmath28 .
( the corresponding region with @xmath1004 , can be treated with a symmetry argument as in the end of section [ sectionsk ] , and this argument will be given later . )
it follows from ( [ tr.12 ] ) that here @xmath1005 and @xmath1006 is well - defined , while @xmath1007 may have zeros .
consequently we use the reflection identity , to get @xmath1008 _ now using ( [ sk.1 ] ) we get @xmath1009 where @xmath1010 where the terms are the same as in ( [ sk.1.5 ] ) , although we shall now use different approximations . again we introduce @xmath0 times the real parts of the different exponents : @xmath1011 as before , we have again , we define the different curves @xmath732 by @xmath742 , for @xmath1012 with the exception of @xmath1013 and @xmath1014 . (
the segment @xmath859 is now defined to be the segment of the positive imaginary axis joining 0 to the lower part of @xmath852 , provided that this lower part is not hidden in the forbidden region , in which case we let @xmath859 be empty ) .
more explicitly , we get : with the function @xmath730 now differs from that of section [ sectionsk ] by a term @xmath1015 . the definition of @xmath749 , @xmath1016 coincides with that in section [ sectionsk ] in the overlap region .
we shall also define a set @xmath1017 for @xmath1018 as in the preceding section .
to do so , we check that if in this region , we also have in the region ( [ sk.7 ] ) we can define @xmath1017 by @xmath1019 , and see that we get a curve with @xmath1020 . using ( [ sk.8 ] )
, we also see that if we represent @xmath1021 by @xmath1022 , then actually the upper bound can here be improved to @xmath1023 . in analogy with section [ sectionsk ] ,
we define a diamond shaped of each zero @xmath891 of @xmath692 by the previously defined @xmath1017 can hit at most one of the @xmath1024 and if that happens , we add that diamond to the set @xmath1017 .
now define the skeleton as before : and as before we can describe the regions of dominance : @xmath692 dominates at distance @xmath1025 below @xmath1026 , for @xmath690 , intersected with the complement of the union of the diamonds .
the other @xmath571 dominate according to the earlier rules in their respective regions at a distance @xmath1027 from the skeleton .
we now consider the case when @xmath136 belongs to the set where @xmath1028 , @xmath28 . from ( [ tr.12 ] )
, we get @xmath1029 @xmath1030 where @xmath630 is non - vanishing , while @xmath1031 may have zeros , so we use the reflection identity to write @xmath1032 we then use ( [ sk.1 ] ) to get @xmath1033 with @xmath1034 @xmath1035,\ ] ] @xmath1036,\quad a_3={\mbox{\rm exp\,}}\left[{i\over h}s_{3,4}+{\pi \mu \over 2h}\right],\ ] ] @xmath1037\ ] ] _ again , we introduce @xmath0 times the real parts of the different exponents : @xmath1038 @xmath1039 @xmath1040 we shall now make the same symmetry transformations as in section 8 , to see that the functions @xmath1041 , @xmath1042 , @xmath1043 , and @xmath1044 play the same role as @xmath1045 , @xmath1043 , @xmath1042 , and @xmath1046 respectively , in the previously considered case : \1 ) add @xmath928 to each of the @xmath1047 .
\2 ) consider the @xmath1047 as s of @xmath929 .
then we get @xmath1048 @xmath1049 @xmath1050 thus apart from a change of sign in @xmath1051 , we see that @xmath1052 has the same properties as @xmath927 in the previously considered case .
_ remark . in the overlap region @xmath1053 where both cases apply ,
we notice that trivially @xmath1054 for @xmath1055 , @xmath1056 , ( and these estimates improve by ( [ sk.44 ] ) when @xmath1057 increases ) . as in the remark at the end of section [ sectionsk],the distance between the two skeletons , defined
according to case 1 and according to case 2 , is therefore @xmath1058 . _
in each of the cases 1 and 2 of sections [ sectionsk ] , [ sectionsk ] , we defined a skeleton @xmath1059 consisting of a horizontal part @xmath852 possibly with a vertical part ( @xmath859 in case 1 and @xmath1060 in case 2 ) added .
we notice that the definitions in the two sections agree for each of the cases 1 and 2 in the overlap regions for the two sections , and we saw in the remarks at the end of the sections , that if we compare the skeletons for the two cases in the overlap region @xmath1061 then the distance between the corresponding skeletons is @xmath1062 now define the body by widening the skeleton : here @xmath1063 may be empty and will now be defined .
they are non - empty if @xmath852 stays entirely in the admissible regions for one of the cases , and in order to fix the ideas , we assume that this is case 1 , and @xmath852 does not intersect the negative imaginary half axis . if so , we have a non - empty segment @xmath859 in the imaginary axis , joining 0 to the closest imaginary point of @xmath852 .
recall that we have defined the diamonds @xmath893 around the zeros of @xmath692 in @xmath859 , by ( [ sk.31],[sk.11 ] ) .
we define @xmath1064 to be the union of @xmath1065 ( in case 1 , and @xmath1060 in case 2 ) and the corresponding diamonds .
( in case 2 we do the corresponding definition with `` 4 '' replaced by `` 1 '' and @xmath1060 is then the segment in the negative imaginary axis , joining 0 to the closest part of @xmath852 . )
@xmath1066 is non - empty precisely when @xmath1065 or @xmath1060 is . in case 1
, it is defined to be the set of points @xmath136 below @xmath852 at distance at most @xmath1067 from @xmath852 with @xmath28 large and with @xmath1068 . here
the upper bound @xmath0 in the last estimate , may be replaced by @xmath1069 for any fixed @xmath108 , and we could decrease @xmath1066 further by a more detailed discussion . in case 2
we have the analogous definition .
we next define what we mean by an _ admissible curve .
it should be a piecewise @xmath636-curve @xmath1070\to{\bf c}$ ] without self - intersections , parametrized by arc - length .
it is tacitly assumed that we consider a family of such curves , which is in the sense that we have uniform bounds on the number of jump discontinuities of @xmath1071 , the continuity of @xmath1071 between the discontinuities , and on the length @xmath1072 .
it is also required that @xmath1073 may belong to @xmath251 only for @xmath1074 , @xmath1075 , where @xmath1076 are disjoint intervals of length @xmath1077 , for some @xmath1078 , if @xmath1079 and of length @xmath1080 otherwise .
we also assume that we have a uniform bound on the number @xmath13 of such intervals .
_ assume for simplicity that @xmath1081 and let us partition @xmath1082 $ ] into intervals in increasing order : for each @xmath1083 , let @xmath1084 , be the corresponding dominant term along @xmath1085 . for simplicity we shall assume that the image of @xmath658 is entirely contained in the admissible region for one of the cases 1 or 2 , so that @xmath1086 is either in @xmath1087 ( case 1 ) , or in @xmath1088 ( case 2 ) .
let @xmath1089 and put @xmath1090 then we have , with @xmath1091 ( cf .
( [ sk.1.5 ] ) , ( [ sk.41 ] ) , ( [ sk.1 ] ) ) : [ thev.1 ]
let @xmath658 be an admissible curve as above .
then @xmath1092 where the maximum is taken over all @xmath140 with @xmath1093 , so if @xmath658 never meets @xmath1066 , we only have the remainder @xmath654 .
notice that the first term of the of ( [ ev.3 ] ) can also be written consider an interval @xmath1083 .
if we first assume @xmath1094 ( if we are in case 1 ) , then for @xmath1095 : let @xmath1096 and @xmath1097 be the start and the end points of @xmath1098 , so that @xmath1099 with the @xmath1100 improving to @xmath1101 if the corresponding neighboring interval @xmath1102 does not meet @xmath1066
. using ( [ ev.5 ] ) , we see that in the case @xmath1103 , we know that @xmath692 is dominating along @xmath1098 and ( [ ev.7 ] ) holds with @xmath1104 replaced by @xmath1105 .
now we also know that along @xmath1098 , we have @xmath1106 with @xmath1107 and with @xmath1108 of variation .
it follows that the real part of the equation ( [ ev.7 ] ) still holds in this case .
we next estimate the integral along @xmath1109 , and let us consider the worst case , when @xmath1110 .
let @xmath1111 be such that @xmath1112 on this disc , we write where @xmath116 is , and @xmath1113={\mbox{\rm exp\,}}{\cal o}(\ln\ln { 1\over \langle \mu \rangle _ h}).&\nonumber\end{aligned}\ ] ] using ( [ ev.9 ] ) and the elementary arguments recalled in the second part of section [ sectionze ] , we get on the other hand , combining the real parts of ( [ ev.7 ] ) , ( [ ev.8 ] ) , ( [ ev.10 ] ) ( [ ev.11 ] ) , we get @xmath1114 with the remainder improving to @xmath654 if we do not encounter @xmath1066 .
now , @xmath1115 , @xmath1116 , and @xmath1117 with the last remainder improving to @xmath1118 , if we avoid @xmath1066 , and ( [ ev.3 ] ) follows .
we end this section by some rough estimates on the location of the skeleton and the corresponding distribution of s for the reduced s constructed in sections [ sectionred ] , [ sectionexp ] .
our starting point is the reduced symbol in ( [ red.8 ] ) and the corresponding 1-dimensional symbol here we shall take @xmath207 real ( and eventually of the form @xmath1119 ) .
if @xmath537 is the original spectral parameter , we introduce the new spectral parameter @xmath129 , by and we will work under the assumption @xmath2 . recall the normal form for @xmath153 near the branch point , given by proposition [ prop1d2 ] and in particular ( [ 1d.31 ] ) : where the leading symbol in @xmath1120 is @xmath524 with @xmath1121 , given in proposition [ prop1d1 ] .
correspondingly , we replace @xmath129 by the new spectral parameter @xmath136 , given by we next estimate the location of the skeleton in the @xmath136-plane , and start with the case @xmath884 .
assume for simplicity that we are in the case 1 : @xmath1122 .
we will only be concerned with the horizontal part @xmath852 of the skeleton .
when @xmath1123 , it is given by the curves @xmath725 , @xmath1124 in ( [ sk.4 ] ) , where @xmath1125 clearly @xmath1126 is lipschitz continuous and for @xmath750 , we get @xmath1127 according to ( [ as.4.1 ] ) , we have @xmath1128 . when @xmath577 , @xmath578 , we know that the leading part of @xmath153 in ( [ ev.13 ] ) is real - valued ( assuming that @xmath57 is real for simplicity ) , so it follows in this case that when @xmath136 is real , then @xmath1129 . since @xmath676 depends ally on @xmath136 , we conclude that in general now combine this with ( [ sk.4 ] ) , the estimate @xmath1130 and proposition [ propa3.1 ] to conclude that in the region @xmath884 , @xmath736 , the horizontal part @xmath852 of the spectrum is given by the union of two curves of the form with @xmath1131 satisfying ( [ a3.21 ] ) , and further , in the left half - plane , we recall that @xmath852 has a more complicated structure .
assume , to fix the ideas , that @xmath828 .
then @xmath852 is the union of the curves ( defined in ( [ sk.5 ] ) ) : @xmath1132 in the region @xmath838 . here
@xmath1133 and the two curves cross at @xmath810 . in the region
@xmath854 @xmath852 is given by @xmath1134 and this curve also contains @xmath810 .
when @xmath825 , we have @xmath1135 so again @xmath1136 satisfies ( [ ev.19 ] ) , while @xmath1137 for @xmath838 , where @xmath1138 satisfy ( [ ev.19 ] ) . in the region @xmath1139
the horizontal part of the spectrum is a union of curves @xmath732 given in ( [ sk.4 ] ) , ( [ sk.5 ] ) . here
the new @xmath730 is lipschitz and @xmath1140 , so the skeleton is here contained in a region the _ overall conclusion is that the skeleton is contained in a region where we recall that @xmath1141 _
we end this section by establishing a simplified statement , to be used in theorem 1.1 .
we shall simply remove a small rectangle around @xmath942 where we have seen that the description of the spectrum is more intricate .
start by recalling the definition of @xmath810 , @xmath812 prior to ( [ sk.18 ] ) .
for instance @xmath810 is the intersection of the curves @xmath250 : @xmath1142 , and @xmath807 : @xmath1143 , where @xmath1144 satisfies ( [ ev.19 ] ) . using that @xmath1145
, we get @xmath1146 implying , we have of course the same estimates for @xmath812 .
choose @xmath28 large so that the `` black box '' contains @xmath810 , @xmath812 .
then we have [ propev.2 ] the number of s in @xmath1147 is @xmath1148 .
the s outside @xmath1147 are exponentially close to @xmath1149 .
more precisely introduce @xmath1150 then there is a bijection @xmath865 ( possibly after a slight modification of @xmath1151 ) between the set of s outside @xmath1147 and @xmath1152 , such that @xmath1153 .
we may first notice that we can replace the index @xmath1154 by @xmath1155 without changing the validity of the statement of the proposition , since @xmath1156 , @xmath1157 .
in view of ( [ ev.21 ] ) , we know that there are no s outside @xmath1151 with @xmath1158 and the discussion in section [ sectionsk ] then shows that the s outside @xmath1147 have to be exponentially close to @xmath1159 and that there is a bijection @xmath865 as stated . to estimate the number of s inside @xmath1147
, we simply apply theorem [ thev.1 ] . with @xmath658
a rectangular contour containing @xmath1147 but contained in @xmath1160 and working directly with @xmath692 instead of @xmath688 .
consider @xmath1161 of the preceding proposition . in view of ( [ sk.1.5 ] ) , it is given by the quantization condition here we recall from the beginning of section [ sectionas ] , that the term @xmath1162 is @xmath1163 , as @xmath1164 .
we also know that if @xmath1165 denote the small additional parameters in the , then @xmath1166 , for @xmath1167 , where @xmath1168 are smooth in @xmath404 and in @xmath519 .
hence the condition takes the form where in the space of s defined in truncated sector : @xmath1169 with and notice that the singularity structure of @xmath1170 is essentially unchanged if we replace @xmath136 by @xmath1171 , where @xmath1172 are classical symbols of order 0 in @xmath0 with s that are near @xmath942 , @xmath1173 and with @xmath229 elliptic , @xmath1174 , @xmath1175 . on the other hand , in the region @xmath1176 , @xmath1177
, we know ( and that was done for instance in @xcite ) , that the s sit on a curve and are given by a bohr sommerfeld condition where @xmath1178 is a classical analytic symbol of order 0 : @xmath1179 , and where here @xmath1180 denotes a closed loop in the energy surface @xmath1181 with @xmath129 and @xmath136 related by , that can be obtained from the real energy curve we get by taking @xmath136 real and putting @xmath1173 .
clearly @xmath1182 , so our discussion gives detailed description about how one can push the standard wkb - construction to the limit @xmath675 in the region @xmath706 .
the same discussion applies to @xmath1183 , @xmath1184 .
we get the conditions where @xmath1185 , @xmath1186 are defined in the truncated sector @xmath789 , @xmath1187 , and @xmath1188 have the analogous properties to those in , , for @xmath1189 , while the first part of should be replaced by the condition that @xmath1190 is the action along a closed loop inside the appropriate complex energy curve , that can be obtained by deformation from the case @xmath750 , @xmath1173 where we take the left real component , close to the left loop in the @xmath68-shaped set @xmath91 . for @xmath1191
we deform from the right real component .
_ one dimensional grushin s. we may assume here without loss of generality , that @xmath57 is real - valued . then we know that @xmath620 in ( [ 1d.33 ] ) satisfies where @xmath1192 is real - valued when @xmath1193 , @xmath1194 .
recall that @xmath620 and @xmath1195 depend ally on @xmath207 and smoothly on @xmath1196 .
_ also recall that the spectrum of @xmath153 is localized to the region ( [ 1d.8 ] ) : @xmath1197 ( as follows also from the more refined estimate ( [ ev.21 ] ) and in view of ( [ 1d.34 ] ) : @xmath1198 , it follows that @xmath136 is localized to a domain of the same type .
we shall now introduce three different grushin s for @xmath1199 in the spirit of @xcite , @xcite .
let @xmath1200 be equal to one near ( 0,0 ) .
we realize @xmath533 as an @xmath0- , that we also denote by @xmath533 , using a gaussian resolution of the identity ( see @xcite , @xcite ) , so that our calculus errors will be exponentially small rather than just @xmath1201 .
assume , in order to fix the ideas , that the support of @xmath1202 is a thin annulus around ( 0,0 ) , containing the points @xmath552 , @xmath432 ( see section [ section1d ] ) .
recall the definition of @xmath559 in ( [ 1d.38 ] ) .
let @xmath1203 be the corresponding s defined for @xmath1204 , depending anti - ally on @xmath129 .
define here @xmath1205 is a small of @xmath552 and @xmath1206 where @xmath1207 is equal to 1 near @xmath552 and we also let @xmath1208 denote the corresponding gaussian quantization .
we may normalize the choice of @xmath1203 so that expanding the commutator , we see that the definition of @xmath1209 ( up to an exponentially small error ) is of the choice of @xmath533 , provided that @xmath273 is a null - solution of @xmath1199 .
assume for simplicity that @xmath533 is real - valued and that the corresponding quantization is .
put our first grushin is directly adapted to the derivation of the quantization condition ( [ 1d.49 ] ) in section [ section1d ] .
it is the form with using section [ section1d ] we see as in @xcite that it is well posed for @xmath129 in some fixed complex of @xmath88 with a solution of the form here we get where the parenthesis is the same as in the quantization condition ( [ 1d.49 ] ) . as usual
, we read off the approximate s as the zeros of e@xmath1210 .
the drawback with this first is that the solution ( [ gr.6 ] ) will grow exponentially when @xmath750 .
this can be seen either directly from the explicit formulae for @xmath1211 and the slightly less explicit expression for @xmath1212 , or from the fact that for @xmath750 , we have approximately a double well problem and with @xmath1213 we prescribe the solution @xmath273 in ( [ gr.4 ] ) in one of the wells , and hence @xmath273 will in general be exponentially large in the other well . of course
, we will have to accept some exponential growth with a rate @xmath1214 but certainly would like to avoid exponential growth with a fixed rate when @xmath136 is real .
it seems impossible to cover a full of @xmath1215 with a single whose solution does not exhibit exponential growth in some region , so we shall use 2 s where one will be nice roughly in the upper half - plane and the other in the lower half plane . the second is designed to cover the region @xmath1216 with some margin .
it is of the form ( [ gr.4 ] ) with so that @xmath1217 , @xmath1218 . for the corresponding model for @xmath445 in section [ sectiontr ] , we get ( cf([tr.11 ] ) , ( [ tr.12 ] ) : this is basically the approach of @xcite and as there , we see by direct calculation or by a more general normalization argument that @xmath1219 is unitary when @xmath136 is real .
we also see that @xmath210 is in any disc @xmath1220 for for @xmath136 outside an angle around @xmath1221 with @xmath1222 , we apply stirling s formula ( [ as.2_- ] ) and get with the @xmath133 sign valid for @xmath1223 , @xmath1224 , and the @xmath131 sign for @xmath1225 , @xmath1226 . as in section [ sectionsk ] , we get we now also assume that @xmath1227 , so that @xmath1228
. we shall estimate the exponent from above .
, the worst exponent is the one with @xmath1229 in the middle term and we approximate @xmath1230 which is dominated by the first term and hence when @xmath736 . when @xmath794 , the worst case is the one with @xmath1231 in the middle term and we approximate @xmath1232 leading to ( [ gr.14 ] ) also in this case . we conclude that @xmath210 is in a domain of the form as a consequence , we get [ propgr1 ] the with @xmath1233 given by ( [ gr.8 ] ) is ly well - posed ( with errors @xmath1234 for @xmath1235 small with @xmath136 in @xmath1236 or away from a small conic of @xmath1237 . if we write the solution as in ( [ gr.6 ] ) , then for @xmath136 as in ( [ gr.15 ] ) , we have let us also compute @xmath1238 . near the branching point , we recall that we have the relation ( [ 1d.41 ] ) for null - solutions to @xmath532 , equal to @xmath1239 near @xmath552 . to determine @xmath1238 , we consider ( [ gr.4 ] ) with @xmath1240 , so that @xmath1241 for @xmath570 .
we then want to express @xmath1242 , @xmath566 in terms of @xmath1243 .
we can do this using ( [ gr.9 ] ) , ( [ gr.10 ] ) above and redo some of the work in section [ section1d ] , but it is easier to to use the work already done and `` solve '' ( [ 1d.41 ] ) .
we get notice that @xmath1244 , @xmath1245 in our ( with @xmath1240 ) are the discontinuities we obtain at @xmath608 , @xmath589 when trying to extend a null - solution near ( 0,0 ) with prescribed @xmath1246 near @xmath1247 , @xmath570 , to a global null - solution near @xmath91 .
we get where it follows that @xmath1248 from the middle expression , we see that this determinant is equal to the expression in the quantization condition ( [ 1d.49 ] ) times a non - vanishing factor .
the third is designed to cover the region @xmath1249 .
it is of the form ( [ gr.4 ] ) with for the corresponding model for @xmath445 in section [ sectiontr ] we get ( cf ( [ tr.11 ] ) , ( [ tr.12 ] ) ) : with @xmath210 as in ( [ gr.9 ] ) , unitary for real @xmath136 , so that thus from ( [ gr.14 ] ) we see that the matrix elements @xmath1250 satisfies and @xmath1251 is in a domain of the form [ propgr2 ] the with @xmath1233 given by ( [ gr.21 ] ) is ly well - posed ( with errors @xmath1252 ) for @xmath1235 small with @xmath136 in @xmath1236 or away from a small conic of @xmath467 .
if we write the solution as in ( [ gr.6 ] ) , then for @xmath136 as in ( [ gr.25 ] ) , we have the estimate ( [ gr.16 ] ) .
we now compute the corresponding @xmath1238 so we put @xmath1240 in ( [ gr.4 ] ) and repeat the arguments above .
near the branching point @xmath273 is a null - solution , @xmath1253 near @xmath552 , now with @xmath1254 , @xmath566 . by `` solving '' ( [ 1d.41 ] )
, we express @xmath1243 in terms of @xmath1255 and find we then get and @xmath1256 _ the global grushin .
we first explain what the natural range will be for @xmath11 .
our global grushin will be built from a direct sum of s for the s @xmath1257 in ( [ red.5 ] ) .
these s can be written where @xmath153 is the appearing earlier in this section and in section [ section1d ] . _ clearly , we add conditions `` @xmath1233 '' only for such @xmath140 for which @xmath1258 is close the spectrum of @xmath153 , i.e. for which ( cf .
( [ 1d.8 ] ) ) , assuming for simplicity that @xmath57 is real - valued .
then according to ( [ gr.16 ] ) we can expect that our will have an inverse @xmath1259 with norm @xmath1260 .
now we can not expect to have a complete decomposition into a direct sum of s , but proposition [ propexp3 ] shows that so is possible up to an error @xmath1261 .
thus in order to absorb the error by a standard perturbation argument , we need @xmath1262 which would follow from or equivalently @xmath1263 here we already know that @xmath1264 , since we assume @xmath1265 , so the new constraint is @xmath1266 . from now on we work in the range since @xmath164 is real - valued with @xmath316 , it follows from ( or in the general case , when @xmath57 is not assumed to be real ) that the s have disjoint spectra .
when is never satisfied for any @xmath140 , it is straight forward to see that @xmath537 is not in the spectrum of our original @xmath39 .
assume now that holds for ( at most ) one @xmath1267 .
let @xmath1268 be the corresponding s @xmath1233 defined earlier in this section . using the notation of proposition [ propred1 ]
, we define @xmath1269 for @xmath1270 , @xmath1271 . here
@xmath1272 . repeating the arguments from @xcite
we see that we get a well - posed with here @xmath1273 is the `` @xmath1238 '' of the approximate 1-dimensional treated in the appropriate one of the propositions [ propgr1 ] , [ propgr2 ] , depending on the sign of the corresponding parameter @xmath843 . in the expressions ,
, we have the term @xmath1274 , where @xmath1275 are given by .
we see that @xmath1276 , so we conclude from this and that the remainder in is @xmath1277 times the dominating term in the expression for @xmath1273 in , or respectively .
_ this implies that if we pass to the @xmath136-variable ( for @xmath1267 ) and define the skeleton as in sections [ sectionsk ] , [ sectionsk ] and the corresponding body @xmath251 as in the beginning of section [ sectionev ] , then the zeros of @xmath1278 are confined to @xmath251 and theorem [ thev.1 ] still applies to give the number of s ( in the @xmath136-plane ) inside an admissible curve . _
we start the discussion in this section with the following general observation .
let @xmath39 be a smooth family of operators , satisfying all the assumptions of the introduction , and in particular ( [ h1 ] ) .
in proposition 2.1 we have seen that the operator @xmath1279 can be reduced by successive averaging procedures to assume that the subprincipal symbol of @xmath1279 vanishes and that the spectrum of @xmath1279 clusters into bands of size @xmath1280 , for some integer @xmath1281
. then @xmath1282 are of @xmath21 for @xmath1283 in ( [ para.1 ] ) .
since the subprincipal symbol vanishes , we already know that @xmath1284 .
suppose that the conclusion of the proposition does not hold and let @xmath1285 be the smallest @xmath667 with @xmath1286 non - constant for some @xmath1287 .
take a of gaussian quasimodes @xmath1288 , @xmath1289 with @xmath1290 see , e.g. @xcite for the standard construction of such quasimodes . then put @xmath1291 with @xmath1292 such that @xmath1293 , so that hence , since we are dealing with s , and varying @xmath172 , so that the values @xmath1294 fill up a whole interval , we get a contradiction to the clustering assumption . .
proposition 12.1 remains to hold in the case described in section 4 of @xcite , where the operator @xmath1279 is conjugated into a normal form in a of a lagrangian torus , rather than near a closed @xmath158trajectory .
from now on we shall assume that @xmath1279 satisfies the assumptions of proposition 12.1 .
let us now switch on @xmath1295 .
an application of proposition 2.1 together with proposition 12.1 then shows that microlocally , near a closed @xmath158trajectory , @xmath201 can be reduced to the form it follows therefore that in the results of @xcite ( theorems 6.4 and 6.7 there ) , we can replace the exponent 2 by the exponent @xmath1296 in the parameter range for @xmath1295 .
thus for the study of the spectrum of @xmath201 in a region where @xmath1297 and @xmath1298 , when @xmath69 is a non - critical value or a maximum or minimum of @xmath1299 along @xmath176 , it suffices to assume that for some @xmath147 . in the case
when @xmath69 is a saddle point value of @xmath1300 , from theorem 1.1 we get the condition indeed , in the latter case we still have proposition [ propexp3 ] and the decoupling condition analogous to becomes @xmath1301 , which is fulfilled by .
we shall now apply these observations to improve the result of proposition 7.1 of @xcite , giving a description of the individual barrier top resonances of the semiclassical schrdinger operator in the resonant case . before doing so , and also for the future use in section 13
, we shall first briefly recall the general setup in section 7 of @xcite , as well as in section 5 of @xcite . as in @xcite ,
let us consider where @xmath1251 satisfies the general assumptions of section 7 in @xcite , allowing us to define the resonances of @xmath321 in the lower half - plane inside some fixed of @xmath1302 , where @xmath1303 , @xmath1304 , @xmath1305 .
as in @xcite , we assume that @xmath1306 is the only trapped @xmath1307-trajectory in @xmath1308 . after a linear symplectic change of coordinates , we may write where @xmath1309 and @xmath1310 is a homogeneous of degree @xmath1311 .
recall further from @xcite that the study of s of @xmath321 near @xmath1312 can be reduced to an for @xmath321 after applying some variant of the method of complex scaling , and that near @xmath30 this simply amounts to working in the new real coordinates @xmath1313 , given by @xmath1314 , @xmath1315 . performing the scaling and dropping the tildes from the notation
, we see that the problem reduces to studying the s close to @xmath88 of the @xmath1316 , now elliptic outside a small of @xmath79 , with symbol where here we continue to write @xmath321 to denote the scaled operator .
we assume that @xmath1309 in ( [ bt.03 ] ) are rationally dependent , which implies that the @xmath1317-flow is periodic . as in @xcite , @xcite ,
we are interested in eigenvalues @xmath145 of @xmath321 with @xmath1318 , @xmath1319 . after a rescaling @xmath1320 , and dropping the tildes over the new variables , we get an operator @xmath1321 that we view as an @xmath1322-pseudodifferential operator with the symbol @xmath1323 here @xmath1324 . now
the spectrum of @xmath1279 is that of the harmonic oscillator , and hence it clusters into sets of diameter @xmath88 and separation of order @xmath0 .
an application of proposition 12.1 shows that all the @xmath1325 in ( [ para.1 ] ) are constant .
moreover , since in this case all the s depend linearly on @xmath0 , we see from the proof that the @xmath1325 have to vanish .
it follows from ( [ para.6 ] ) that in the zone corresponding to non - critical values @xmath69 or non - degenerate maxima or minima , the range of energies that we get is [ range1 ] h^2n_0/(1 + 2n_0)h^2/(1 + 2 ) , for all @xmath1326 and all @xmath1327 .
when @xmath69 corresponds to a branching level , we get from ( [ para.7 ] ) [ range2 ] h^2n_0/(1 + 2n_0)h^1/2 .
we summarize the discussion above in the following proposition , which is an improvement of proposition 7.3 in @xcite .
clearly , in a similar fashion , we also obtain an improvement of theorem 5.1 in @xcite .
assume that the principal symbol @xmath1328 in ( [ bt.00 ] ) has an asymptotic expansion ( [ bt.01 ] ) , and assume that ( [ bt.2 ] ) holds .
assume furthermore that the function @xmath1329 , defined as the average of @xmath1330 along the hamilton flow of @xmath35 in ( [ bt.1 ] ) does not vanish identically .
then the resonances of the operator @xmath321 in the domain [ para5 ] \ { z ; h^2n_0/(1 + 2n_0)h^ } \\{z ; < ^3/2 } , where @xmath1331 , @xmath1327 , and @xmath1326 are arbitrary but fixed , are given by ~e_0 - i(h(k_1-/4)+^3 _
j=0^ h^j ^-2j r_j((k-)- , , ) ) , with @xmath1332 we have @xmath1333 , @xmath1334 with @xmath1335 , and @xmath1336 is fixed , and we choose @xmath1337 with @xmath1338 .
the union in ( [ para5 ] ) is taken over the set of critical values of @xmath1329 , restricted to @xmath1339 , with @xmath250 varying over this set . .
if @xmath1329 restricted to @xmath1339 has precisely one non - degenerate saddle point with the critical value @xmath250 , then the results of the present paper apply and give a description of the individual resonances in a half - cubic neighborhood of the curve @xmath1340 . in the following section , we shall consider explicit examples of homogeneous polynomials for which the assumptions of theorem 1.1 are satisfied .
this section is devoted to a study of examples of potentials of the schrdinger operator ( [ bt.00 ] ) to which theorem 1.1 is applicable .
let us recall from ( [ bt.03 ] ) that we are interested in eigenvalues close to @xmath88 of the operator @xmath321 , elliptic outside a small of @xmath79 with symbol [ bt1 ] p(x,)=p(x,)+ie^p_3(x)+ie^i p_4(x)+ , where the harmonic oscillator @xmath1341 has been defined in ( [ bt.1 ] ) . as before
, we make the resonant assumption ( [ bt.2 ] ) .
consider first a general perturbation of @xmath35 of the form of a linear combination of terms @xmath1342 with @xmath1343 , for some @xmath1344 .
recall from @xcite how to compute the corresponding trajectory average @xmath1345 : basically we use action - angle coordinates , but to start with , we can do things a little easier by introducing and notice that along a @xmath158-trajectory we get in the @xmath1346 coordinates : then write @xmath1347 , @xmath1348 , so that @xmath1349 then expand the product by means of the binomial theorem .
the time average is equal to the time- term and since this average is constant along each trajectory we shall replace the symbols @xmath1350 simply by @xmath1351 .
in this section we consider the case when @xmath1352 , @xmath1353 , @xmath1354 .
( in @xcite we noticed that in this case the average will vanish when @xmath1355 and in @xcite we made a more refined study of that case taking into account one more term in the perturbative expansion ) .
this means that we take @xmath1356 , and for simplicity we also assume that @xmath1357 for all odd @xmath24 in ( [ bt1 ] ) , so that we can concentrate on the perturbation @xmath1358 in ( [ bt1 ] ) . performing a rescaling as described in the previous section , with @xmath1295 replaced by @xmath1359 , ( i.e. setting @xmath1360 rather than @xmath1320 ) , we get and here , as before , we choose @xmath1 of the same order of magnitude as the modulus of the s for the @xmath1361 , that we want to study .
now we continue the calculations of trajectory averages using ( [ bt.5 ] ) .
we have @xmath1362 in the same way , we get next look at the averages of mixed terms : @xmath1363 @xmath1364 notice that our averages are invariant under the anti - symplectic involution this is necessarily the case since we stay in the framework of ordinary schrdinger operators ( without magnetic fields ) whose symbols have this invariance .
now write our results in the action angle variables @xmath1365 , given by so that @xmath1366 it follows from the hamilton equations that @xmath542 and @xmath1367 are constant along every @xmath158-trajectory .
the involution @xmath641 can also be described as @xmath1368 , and hence in action
angle variables as we shall study our averages as s on the abstract symplectic using the @xmath1369-coordinates , we have and the equivalence relation induced by the @xmath158-flow is : @xmath1370 if and only if @xmath1371 for some @xmath184 .
thus we see that @xmath1372 can be identified with the complex projective space @xmath1373 .
it is well known that this space is diffeomorphic to @xmath1374 .
indeed , @xmath1373 can be identified with the 1-point compactification @xmath1375 via the map @xmath1376 and the one point compactification can be identified with the riemann sphere .
@xmath1372 can be parametrized by @xmath1377 with @xmath1378 , @xmath1379 , @xmath1380 , with the convention that all the @xmath1381 denote the same point and similarly for @xmath1382 .
the involution @xmath641 induces the anti - symplectic involution notice that the set of fixed points of @xmath641 is given by all points with @xmath1383 or @xmath1384 .
these points form a ( great ) circle on @xmath1372 and can also be described as the set of trajectories in @xmath1339 whose @xmath387-space projections hit the of the potential well @xmath1385 .
we consider perturbations of the form then on @xmath1372 we get with @xmath1386 , so that @xmath1387 : @xmath1388 where we used that @xmath1389 . we are interested in the critical points of this on @xmath1372 , and the values @xmath1390 will have to be treated separately .
in particular we are interested in the number of saddle points . if we have only one saddle point we will be able to apply the results of this paper .
this is still the case if there are two saddle points provided that the corresponding critical values are different .
we will also encounter the case of two saddle points @xmath1391 away from the equator and then necessarily with @xmath1392 . in that case
the critical vales will be equal and the results of this paper will not apply directly .
we plan to return to that case in a future paper , where the role of symmetries will be studied .
put @xmath1393 then , notice that @xmath1394 is critical precisely when @xmath1395 and that @xmath1396 $ ] .
when @xmath1397 , we may treat @xmath164 as an variable .
the same observation is valid for @xmath13980,{1\over 2}]$ ] .
it is non - critical in @xmath1399 ( i.e. for @xmath1400 ) .
( as already mentioned , the value @xmath1401 , corresponding to @xmath1390 , will require a different treatment . ) in order to avoid various degenerations , we shall assume \1 ) _ critical points with _ here both @xmath1402 and @xmath164 can be treated as variables and the critical points are determined by @xmath1403 this can also be written @xmath1404 under the assumption ( [ bt.23 ] ) we have @xmath1405 , so we get @xmath1406 this is in contradiction with the assumption ( [ bt.22.5 ] ) , so we conclude that there are no critical points away from the union of the `` vertical circle '' given by @xmath1407 and the horizontal circle : @xmath1408 .
\2 ) _ critical points on the horizontal circle away from the vertical one : _ then @xmath1409 and
this is a critical value , so we only have to look for critical points @xmath1402 , leading to @xmath1410 recall that @xmath1411 under the assumption ( [ bt.24 ] ) , so we reach the conclusion that if then there are two distinct critical points in the region ( [ bt.24 ] ) , given by and otherwise there are no such points .
( in the remaining degenerate case @xmath1412 the whole horizontal circle is critical . )
we also study the nature of the critical points , by computing the hessian of @xmath57 @xmath1413 . using that @xmath1414 , @xmath1415
, we get at both points so both critical points are of signature @xmath1416 where the first component corresponds to the horizontal ( @xmath495 ) direction .
( we use the convention that a signature described by @xmath1417 is given by @xmath1418 . )
\3 ) _ critical points on the vertical circle away from the horizontal one and from the poles @xmath1390 : _ for @xmath1383 , we have @xmath1419 and we look for critical points of @xmath1420 , leading to hence we get two critical points in this region if and otherwise no point on this half of the vertical circle .
( in the degenerate case @xmath1421 , @xmath1422 the whole vertical circle is critical . ) for @xmath1384 , we have @xmath1423 and we look for critical points of @xmath1424 , leading to so we get 2 critical points in this case if and otherwise no critical points on this half of the vertical circle .
we will see shortly that we have critical points at the poles when @xmath1422 . in both subcases , we get by a straight forward calculation : @xmath1425 so the signature is where the first component refers to the direction of the vertical circle through the critical point .
\4 ) _ the two points of intersection of the two circles : _ here @xmath1408 and @xmath1407 . here
both @xmath164 and @xmath1402 are critical , so our intersection points are both critical . by straight forward calculation , we get for @xmath1426 @xmath1383 , @xmath1408 and for @xmath1427 @xmath1384 , @xmath1408 in particular , the signature is also notice that \5 ) _ it remains to study the `` poles '' , given by _ @xmath1390 . here
the @xmath1428 coordinates degenerate and we return to the @xmath537-coordinates . using that @xmath1429 we get @xmath1430
make the change of variables @xmath1431 and restrict to the energy surface @xmath1339 , which now becomes @xmath1432 .
then we get again with @xmath1433 .
recall that we work on the projective space , described as the 3-sphere @xmath1434 modulo the action of the rotations @xmath1435 .
consider the case @xmath1436 .
correspondingly , we can choose the point @xmath1437 .
the @xmath158-integral curve through that point is @xmath1438 and locally , we can identify @xmath1372 with the transversal hypersurface @xmath175 in the 3-sphere which is given by @xmath1439 .
thus @xmath1440 with @xmath1441 , and we get @xmath1442 we can use the real and imaginary parts of @xmath1443 as local coordinates on @xmath175 . then on @xmath175 , we get the taylor expansion : @xmath1444 we conclude that the `` pole '' @xmath1436 is a critical point iff @xmath1422 and when this point is critical , the signature is @xmath1445 , where the first component corresponds to the direction of the ( vertical ) circle through the pole . by symmetry in the indices 1 , 2
, we have the identical conclusion for the opposite pole , given by @xmath1446 .
notice finally that this case together with the case 3 give a complete description of the critical points on the vertical circle away from the crossings with the horizontal one .
we observe that the critical points away from the intersection of the two circles are non - degenerate and keep constant signatures under small perturbations of the parameters ( except in the degenerate cases @xmath1447 and @xmath1448 ) .
these critical points can only be killed or born by passing through one of the two crossing points .
this happens in the following 4 cases : \1 ) @xmath1449 : the critical points on the horizontal circle coalesce into the crossing point @xmath1450 .
when @xmath1451 goes from @xmath1452 to @xmath1453 , the two critical points disappear and the signature at @xmath1450 goes from @xmath1454 to @xmath1455 \2 ) @xmath1456 .
the two critical points on the horizontal great circle coalesce into the crossing point @xmath1457 .
when @xmath1458 goes from @xmath1459 to @xmath1460 , the signature of that crossing point goes from @xmath1454 to @xmath1461 .
\3 ) @xmath1462 : the two critical points on the vertical circle coalesce into the crossing point @xmath1463 @xmath1464 .
when @xmath1465 goes from @xmath1452 to @xmath1453 , the signature of that crossing point goes from @xmath1466 to @xmath1467 .
\4 ) @xmath1468 : the two critical points on the vertical circle coalesce into the crossing point @xmath1469 .
when @xmath1465 goes from @xmath1459 to @xmath1460 , the signature of that crossing point goes from @xmath1467 to @xmath1466 . in the following
, we may assume in order to fix the ideas , that @xmath1470 .
in the @xmath1471-plane , we define the following open sets , separated from each other by the 4 lines @xmath1472 , @xmath1473 , where all the critical points will be : @xmath1474 .
@xmath1475 .
@xmath1476 .
@xmath1477 .
@xmath1478 .
@xmath1479 .
@xmath1480 .
@xmath1481 .
@xmath1482 .
[ l][l]@xmath1483 [ l][l]@xmath865 [ 1][1]@xmath1484 [ 1][1]@xmath1485 [ 1][1]@xmath1486 [ 1][1]@xmath1487 [ 1][1]@xmath1488 [ 1][1]@xmath250 [ 1][1]@xmath1489 [ 1][1]@xmath1490 [ 1][1]@xmath1491 [ 1][1]@xmath1492 [ 1][1]@xmath1493 [ 1][1]@xmath1494 [ 1][1]@xmath1495 then the earlier discussion gives the location and the signature of the critical points in each of the cases .
let @xmath1496 denote the `` forward '' crossing point of the two circles , given by @xmath1408 , @xmath1497 .
similarly let @xmath1498 denote the `` backward '' crossing point , given by @xmath1408 , @xmath1499 .
@xmath250 : signature at @xmath1496 : @xmath1500 signature at @xmath1498 : @xmath1500 away from the crossings : on the horizontal circle : two critical points with signature @xmath1501 on the vertical circle : two critical points with signature @xmath1502 @xmath1489 : signature at @xmath1496 : @xmath1500 signature at @xmath1498 : @xmath1503 away from the crossings : on the horizontal circle : no critical points on the vertical circle : two critical points with signature @xmath1502 here @xmath1504 is smaller than @xmath1505 but larger than the two other critical values .
@xmath1495 : signature at @xmath1496 : @xmath1503 signature at @xmath1498 : @xmath1500 away from the crossings : on the horizontal circle : no critical points on the vertical circle : two critical points with signature @xmath1502 . here @xmath1506 is smaller than @xmath1507 but larger than the two other critical values .
@xmath1488 : signature at @xmath1496 : @xmath1500 signature at @xmath1498 : @xmath1502 away from the crossings : on the horizontal circle : no critical points on the vertical circle : no critical points @xmath1494 : signature at @xmath1496 : @xmath1502 signature at @xmath1498 : @xmath1500 away from the crossings : on the horizontal circle : no critical points on the vertical circle : no critical points @xmath1490 : signature at @xmath1496 : @xmath1503 signature at @xmath1498 : @xmath1503 away from the crossings : on the horizontal circle : two critical points with signature @xmath1500 on the vertical circle : two critical points with signature @xmath1502 here @xmath1506 , @xmath1504 are larger than the values at the critical points on the vertical circle and smaller than the values at the critical points on the horizontal circle . from ( [ bt.27.85 ] )
we also know that @xmath1508 .
@xmath1491 : signature at @xmath1496 : @xmath1503 signature at @xmath1498 : @xmath1502 away from the crossings : on the horizontal circle : two critical points with signature @xmath1500 . on the vertical circle : no critical points . in this case
@xmath1509 is larger than @xmath1504 but smaller than the two other critical values .
@xmath1493 : signature at @xmath1496 : @xmath1502 signature at @xmath1498 : @xmath1503 away from the crossings : on the horizontal circle : two critical points with signature @xmath1500 . on the vertical circle :
no critical points . in this case
@xmath1510 is larger than @xmath1506 but smaller than the two other critical values .
@xmath1492 : signature at @xmath1496 : @xmath1502 signature at @xmath1498 : @xmath1502 away from the crossings : on the horizontal circle : two critical points with signature @xmath1500 . on the vertical circle : two critical points with signature @xmath1503 . in the cases
@xmath1511 we have precisely one saddle point ( necessarily ) situated on the vertical circle which is the fixed point set of @xmath641 . in these cases ,
the results of this paper apply .
the results also apply in the case @xmath1490 , provided that we assume that @xmath1512 in order to separate the two saddle point values . in these cases
it is easy to understand the structure and the shape of the level sets @xmath1513 .
in particular we see that when we let @xmath861 be a saddle point value we get a connected `` @xmath68 '' shaped set ( and no `` circular '' components on which @xmath57 is non - critical everywhere ) . in the case @xmath1492 , we have two saddle points situated on the vertical circle symmetrically the horizontal circle . since we have chosen to use perturbations which are symmetric under permutation of @xmath1514 , the function @xmath57 is invariant under the map @xmath1515 , so the critical values are necessarily equal . here
we can break the symmetry by adding a small multiple of for instance @xmath1516 so that we still have precisely two saddle points , but with different critical values .
then the results of our paper apply .
in the case @xmath250 , we have two saddle points on the horizontal circle .
they are of course exchanged by application of @xmath641 and this symmetry remains under perturbations within the class of s without magnetic field .
we hope to analyze this case in a future work .
we shall next compute the @xmath984-contribution to the averaging of the principal symbol @xmath1517 appearing in in ( [ bt.5.5 ] ) , by applying the calculations of the end of section [ sectionsk ] , with recall from there , that we have the averaged symbol where @xmath1518 and @xmath1519 were defined in ( [ sk.58 ] ) , ( [ sk.56 ] ) .
a simple calculation gives @xmath1520 more generally , @xmath1521 where @xmath1522 denotes the symplectic form , viewed as an alternate bilinear form on @xmath1523 .
hence @xmath1524 when @xmath1525 , and @xmath1526 otherwise .
if @xmath1527 with @xmath1528 we get by multinomial expansion @xmath1529 when calculating this kind of expressions , it is useful to observe that the relations @xmath1530 , @xmath1531 imply : @xmath1532 , @xmath1533 .
we have the series expansion @xmath1534 combining this with ( [ bt.29.5 ] ) and the parseval identity , we get @xmath1535 using this formula we get after a few days of simple but tedious calculations : now recall that @xmath200 is given by ( [ bt.20 ] ) , so that by ( [ bt.29 ] ) , we have @xmath1536 according to ( [ para.5 ] ) our reduced 1-dimensional has the symbol @xmath1537 put @xmath1538 if we assume that @xmath1539 , then we get according to ( [ sk.50 ] ) : where we recall that @xmath1540 is the ( real ) action along a loop in @xmath1541 starting and ending at the saddle point @xmath1542 , and that @xmath1543 is the corresponding perturbed action for @xmath1544 . from ( [ bt.31 ] ) , we see that @xmath1545 is minimal precisely on the horizontal circle @xmath1546 . in the cases @xmath1547 , @xmath1548 , @xmath1490 the saddle points belong to @xmath1549 situated on that circle . since @xmath57 is invariant under reflection in that circle , either the loop @xmath1550 is entirely in the upper or lower hemisphere intersecting the equator only at @xmath1542 ( and this is happens in the cases @xmath1547 and for one of the saddles in case @xmath1490 ) or @xmath1550 intersects the equator at one more point and is symmetric around the equator ( and this happens in the cases @xmath1548 and for one of the saddles in case @xmath1490 ) . in both cases
we see that taking into account the form of @xmath620 in ( [ bt.37 ] ) we conclude that for every @xmath100 the integral in the of ( [ bt.38 ] ) is @xmath1551 except for @xmath1552 in a set of measure @xmath88 . for @xmath1552 outside that exceptional set
, we conclude from the discussion at the end of section [ sectionsk ] that the spectrum of the one dimensional localized s has a genuinely two - dimensional structure .
to get a complete normal form we shall do further conjugations with analytic s of order 0 in such a way that the complete symbol also becomes a of @xmath519 , @xmath1 , @xmath520 and @xmath521 . moreover , we need to do so with errors that are @xmath1553 ( rather than merely @xmath1201 as in @xcite .
@xmath153 is not a but it has a realization and becomes a , if we allow some of the @xmath0-dependence to appear as an parameter in the s of the @xmath0- expansion .
thus , our starting point will be a symbol of the form where @xmath1554 is and @xmath654 in some fixed complex of @xmath491 , @xmath1555 .
we define the @xmath181-quasi - norm as above , but now it is important that we work in the weyl quantization . to an symbol @xmath229
we then associate the infinite order @xmath1556 as in ( [ exp.05 ] ) . from the definition ,
we verify the following metaplectic invariance property : if @xmath1557 is an affine linear and @xmath1558 , @xmath1559 denote the usual operation of pull - back and push - forward of s on @xmath1560 , then this implies that if we define our quasi - norms with the help of a of opens sets @xmath239 which are invariant under @xmath186 , then in the case of ( [ 1d.13 ] ) , we shall let @xmath239 be of the form @xmath1561 for a suitable @xmath1562 , and we observe that these balls are invariant under @xmath1563 when @xmath970 is real . after applying the inverse of @xmath1564 to our , we may assume that the principal symbol of @xmath153 is @xmath521 , so ( [ 1d.13 ] ) simplifies to with a new @xmath1554 having the same properties as the previous one . using the same letters for s and their symbols , we let @xmath1565 be the quantization of @xmath521 .
notice that @xmath1566 , so that @xmath1567 . if @xmath251 is an analytic @xmath0- of order 0 , we put and notice that on the symbol level , also notice that @xmath1568=0 $ ] . choosing the @xmath181-quasi - norms as above
, we further have that the , has the solution where @xmath1569 is the with support in @xmath1570 $ ] which is affine on @xmath1571 , @xmath530,\pi ] $ ] with @xmath1572 , @xmath1573 .
we have as in section [ sectionexp ] , we see that the map has the differential where under the assumption that @xmath1574
. now return to ( [ 1d.16 ] ) .
after a first conjugation , we may reduce ourselves to the case when @xmath1575 is @xmath1140 , so that for some @xmath28 , when @xmath1576 .
we look for @xmath250 such that @xmath1577 commutes with @xmath955 , and we try @xmath1578 with convergence in some @xmath1579-quasi - norm .
start by solving from ( [ 1d.24 ] ) , ( [ 1d.25 ] ) , we get @xmath1580 where the last estimate holds for @xmath1581 , with @xmath1582 small enough
. choose @xmath1583 with then @xmath1584 with @xmath1585 .
iterating the procedure , we get @xmath1586 with @xmath1587 such that if @xmath1578 , then the previous discussion shows how to find @xmath210 so that modulo an error @xmath1553 , @xmath1588 commutes with @xmath1589 .
moreover @xmath1588 is a classical analytic ( after allowing @xmath0 as an parameter in the s in the expansions ) . put @xmath1590 and work near @xmath1591 for some fixed small @xmath1592 .
then @xmath1593 and since the class of s is conserved under changes of variables , we know that @xmath1594 where @xmath138 is an symbol .
but @xmath1595=0 $ ] , so @xmath1596 and returning to the @xmath387-coordinates , we get the representation ( [ 1d.31]).@xmath1597
for simplicity , we restrict the attention to the right half - plane , @xmath736 and pick one of the equations in ( [ sk.4 ] ) , that we write where @xmath1598 is lipschitz in a of @xmath88 . as we have already observed , so ( [ a3.1 ] ) determines a curve of the form we want to express @xmath620 in terms of @xmath1599 up to small errors .
let us first compare the solution @xmath136 of ( [ a3.1 ] ) with the solution @xmath1600 of the simplified using that @xmath1601 together with ( [ a3.2 ] ) , we see that with this estimate in mind , we now concentrate on the simplified ( [ a3.4 ] ) , and we drop the tildes for simplicity .
assume first that we are in the region then @xmath1602 where we write @xmath1603 .
thus , if @xmath1604 solves ( [ a3.4 ] ) and ( [ a3.6 ] ) holds , then we first see that @xmath1605 and then that so , if we assume then we are in the region ( [ a3.6 ] ) , and the solution @xmath1606 of ( [ a3.4 ] ) takes the form ( [ a3.7 ] ) . combining with ( [ a3.5 ] )
we get under the assumption ( [ a3.8 ] ) , consider first the simplified with @xmath1608 ( both @xmath1609 ) we get try the approximate solution @xmath1610 .
then by a simple calculation , @xmath1611 since the derivative of the in ( [ a3.13 ] ) is close to 1 , we see that the solution @xmath142 of that is of the form @xmath1612 ; hence the solution of ( [ a3.12 ] ) is of the form if we replace @xmath537 by @xmath1613 , we get the order of magnitude of the solution to ( [ a3.11 ] ) : and the assumption that @xmath1614 reads : the earlier arguments show that this condition is equivalent to which indeed is complementary to ( [ a3.8 ] ) . using ( [ a3.16 ] ) , we get @xmath1615 where we notice that @xmath1616 by ( [ a3.17 ] ) .
hence ( [ a3.11 ] ) gives and applying ( [ a3.15 ] ) with @xmath537 equal to the of ( [ a3.19 ] ) , we get @xmath1617 which simplifies to recall that here @xmath1618 in the simplified . to get the corresponding result for ( [ a3.1 ] ) , we apply ( [ a3.1 ] ) and conclude that ( [ a3.20 ] ) holds for @xmath1603 solving ( [ a3.1 ] ) , under the equivalent conditions ( [ a3.17 ] ) , ( [ a3.18 ] ) , and assuming also @xmath1619 , @xmath1620 .
s. beraha , j. kahane , and n. j. weiss , _ limits of zeros of recursively defined families of polynomials _ , studies in foundations and combinatorics , pp . 213232 , adv . in math .
appl . , 1 , academic press , new york
london , 1978 .
m. hager , _ instabilit spectrale semiclassique pour des oprateurs non - autoadjoints i : un modle _ , preprint 2004 , http://hal.ccsd.cnrs.fr/ccsd-00001594 .
annales sci .
de la facult des sciences de toulouse , to appear .
m. hitrik , j. sjstrand , _ non - selfadjoint perturbations of selfadjoint operators in 2 dimensions .
ii . vanishing averages _ , comm .
pde , to appear .
m. hitrik , j. sjstrand , and s. v ngoc , _ diophantine tori and spectral asymptotics for non - selfadjoint operators _ , preprint , 2005 .
see www.arxiv.org/abs/math.sp/0502032 .
l. nedelec , _ perturbations of non self - adjoint sturm - liouville problems , with applications to harmonic oscillators , _ methods and applications . of analysis ,
to appear .
+ see also www.arxiv.org/abs/math.sp/0406491 a. a. shkalikov , _ spectral properties of the orr - sommerfeld with large reynolds numbers _ , j.math . sciences , to appear , translated from sovremennaya matematika .
fundamentalnye napravlenyiya , vol .
3 , differential and functional differential equations , part 3 , 2003 . | this is the third in a series of works devoted to spectral asymptotics for non- perturbations of selfadjoint @xmath0-s in dimension 2 , having a periodic classical flow . assuming that the strength @xmath1 of the perturbation is in the range @xmath2 ( and
may sometimes reach even smaller values ) , we get an description of the s in rectangles @xmath3+i\epsilon [ f_0 - 1/c , f_0 + 1/c]$ ] , @xmath4 , when @xmath5 is a saddle point value of the flow average of the leading perturbation
. # 1([#1 ] ) # 1#2[#1]#2 # 1#2#3@xmath6 # 1#2#3#4@xmath7 # 1#2#3#4#5@xmath8 # 1#2#3#4#5#6@xmath9 # 1#2#3#4#5#6#7@xmath10 # 1#2#1#2 = cmbx10 scaled 1 = cmbx10 scaled 2 = cmsl8 = cmr10 at 8pt = cmr10 at 12pt = cmbx10 at 14pt * keywords and phrases : * non - selfadjoint , eigenvalue , periodic flow , branching singularity , * mathematics subject classification 2000 * : 31c10 , 35p20 , 35q40 , 37j35 , 37j45 , 53d22 , 58j40 |
noncommutative models appear to be relevant to the description of various aspects of string theory @xcite .
an approach to noncommutative gauge theories can be developed using matrix models @xcite in the limit of large matrices .
compactifications of these models to noncommutative tori was shown to yield the noncommutative yang
mills models @xcite .
the noncommutative tori correspond to matrix configurations in the ikkt model having _ nondegenerate _ scalar commutator in the limit @xmath1 .
although , at infinite matrix size these configurations solve equations of motion , they do not correspond to local minima of the classical action , and therefore do not contribute at the level of perturbation theory .
however , due to their large entropical factor these configurations should become important in the strong coupling regime . in the case of finite @xmath2 one can construct a map from the matrix model to some kind of non - commutative lattice gauge model ( for a recent review see @xcite and references therein ) .
the limit @xmath1 was studied by the author in refs .
@xcite , where it was shown that this limit is ambiguous at least in the perturbative approach .
the ambiguity consists in the fact that depending on the background solution chosen one has in the @xmath1 limit either ten - dimensional yang mills type model or its reductions to lower dimensions . in what follows
we will start with a particular choice from the variety of possible models arising in the @xmath1 limit of the ikkt matrix model .
we are going to show that in fact there is a wide universality among these models , in particular , matrix fluctuation around a @xmath0-dimensional ( commutative ) solution are completely equivalent to perturbations around the mentioned configuration having nondegenerate scalar commutator in @xmath3 dimensions and corresponding to noncommutative tori .
this equivalence become manifest due to the possibility to absorb the kinetic term in noncommutative gauge models . also , we explicitly build a one - to - one map from one - dimensional noncommutative gauge model to the @xmath0-dimensional one using the isomorphism of hilbert spaces .
the @xmath0-dimensional model can be implemented as a solution in the one - dimensional case while the last can be obtained through the dimensional reduction of the former .
the plan of the paper is as follows .
first we review some results concerning ikkt model , after that we describe the equivalence between fluctuations around commutative / noncommutative solutions , and finally find the solution which gives the map from two- to the @xmath0-dimensional noncommutative models , and discuss the implications of this map . for the notations and background of this work we refer the reader to the ref . @xcite .
start with the ikkt model at finite @xmath2 .
it is given by the classical euclidean action , @xmath5 ^ 2-\bar{\psi } \gamma^\mu[x_\mu,\psi],\ ] ] where @xmath6 , and @xmath7 are scalar and spinor matrices with large size @xmath1 .
note that in this paper the greek labels always run in ten dimensions , @xmath8 .
this model possesses a number of properties such as supersymmetry , so(10)-lorentz and su(n)-gauge invariances @xcite . at finite @xmath2 ,
the only classical solutions to this model are given by sets of commuting matrices @xmath9 , @xcite , @xmath10=0.\ ] ] if eigenvalues of matrices @xmath9 form a @xmath0-dimensional lattice , where @xmath0 is an arbitrary integer in the range , @xmath11 , then they can be expressed as functions of @xmath0 independent matrices @xmath12 , where the latin indices run through @xmath0 dimensions , @xmath13 . in the limit @xmath1 matrix fluctuations around such a background
are described by a @xmath0-dimensional yang
mills type model .
in particular , when @xmath14 , and matrices @xmath9 are independent one can identify them with @xmath12 .
( in this case the sets of greek and latin indices coincide . )
the limiting @xmath4 model in this case is equivalent to one given by the following action , @xcite , @xmath15 where , @xmath16_{\star}(x , l ) \\ \label{nabla } & \nabla_\mu\psi(x , l)={{\partial}_\mu}\psi(x , l ) -g[{\mathcal{a}}_\mu,\psi]_{\star}(x , l).\end{aligned}\ ] ] here derivatives are computed with respect to @xmath17 , and they are given by , @xmath18,\ ] ] commutators in eqs .
( [ s_func][deriv ] ) are computed using the star product , @xmath19_{\star}={\mathcal{a}}\star { \mathcal{b}}-{\mathcal{b}}\star { \mathcal{a}},\ ] ] while the star product itself is defined as , @xmath20 here @xmath21 ( the same as @xmath22 ) parameterise the spectrum of @xmath12 , while @xmath23 are coordinates on the ( fourier ) dual space @xcite . in the case
when @xmath24 the model is given by the reduction of ( [ s_func ] ) to @xmath0 dimensions .
in this case , @xmath23 and @xmath21 are @xmath0-dimensional and the set of greek indices comes split in two subsets formed by latin indices @xmath25 and other one of kaluza
klein multiplet indices which appear upon reduction to @xmath0 dimensions from 10 dimensions
. we will not introduce the last type of indices , because along this paper we do not need the this split explicitly , but keep instead the greek letters for both space - time and kaluza klein multiplets . in this case
the derivative terms generalise according to ( [ deriv ] ) , and the dimensional reduced action keeps the same form as given by eqs .
( [ s_func][prod- > * ] ) , but one should keep in mind that @xmath26 in this case denote both the @xmath0-dimensional gauge field and the kaluza klein scalars . using the moyal correspondence , the algebra of @xmath27-functions supplied with the star product ( [ scom ] )
can be seen as a representation of the @xmath0-dimensional heiseberg algebra with the weyl ordering prescription which is generated by @xmath21 and @xmath23 , acting on the hilbert space @xmath28 and satisfying the commutation relation , @xmath29=-{\mathrm{i}}\delta_i^j.\ ] ] in this case the integration over @xmath30 is equivalent to taking the trace over @xmath28 . in what follows we will not distinguish between these two forms . by a redefinition of field @xmath31
, @xmath32 one can absorb the kinetic term in eq .
( [ s_func ] ) . as
a result one has the model described by the action ( in heisenberg form ) , @xmath33 ^ 2 + \bar{\psi}\gamma^\mu[{\mathcal{a}}_\mu,\psi]\right),\ ] ] where @xmath34 and @xmath7 are hermitian operators of @xmath0-dimensional heisenberg algebra which act on @xmath35 and are represented by noncommutative functions on @xmath36 , the trace @xmath37 denotes the integration over noncommutative phase space , @xmath38 .
once again note that @xmath39 , while the fields are defined on @xmath3-dimensional noncommutative space generated by @xmath40 and @xmath17 .
let us note that in the form ( [ s_op ] ) the model is manifestly invariant with respect to reparameterisations of @xmath41 preserving the commutator ( [ canon ] ) .
although , the action ( [ s_op ] ) or its dimensional reductions are obtained as a continuum limit of fluctuations around a commutative background ( @xmath42=0 $ ] ) , one can show that the same model can be obtained as a continuum limit of fluctuations around a configuration with @xmath42={\mathrm{i}}b_{\mu\nu}\neq 0 $ ] , in @xmath3 dimensions .
in other words , the model ( [ s_op ] ) is equivalent to the u(1 ) non - commutative yang millls model in @xmath3 dimensions @xcite , when @xmath43 , or its @xmath3 dimensional reduction when @xmath44 .
indeed , consider action ( [ s_op ] ) in the case of @xmath45 .
after shifting back the fields @xmath46 , where @xmath47 are given by @xmath48 for @xmath49 and @xmath50 , for @xmath51 the action ( [ s_op ] ) becomes one of the 10-dimensional noncommutative u(1 ) yang mills model . ) .
this correspondence is possible since @xmath0-dimensional heisenberg algebra coincides with @xmath3-dimensional noncommutative space . ]
the same trick can be made for any @xmath52 , however , the meaning of the model when @xmath53 is not yet clear , but as we are going to demonstrate later this is not a problem since any @xmath54 model is equivalent to the @xmath55 .
in this section we consider the model given by the action ( [ s_op ] ) with @xmath55 , which means that the fields are defined as functions on the one - dimensional heisenberg algebra generated by @xmath40 and @xmath17 , satisfying usual commutation relation , @xmath56=-{\mathrm{i}}\hbar,\ ] ] which is the same as two - dimensional noncommutative space .
let us note that the one - dimensional solution is one having the largest entropical factor @xcite in the ikkt matrix model . for convenience consider the heisenberg algebra to be defined on a circle : @xmath57 .
in this case the momentum operator @xmath40 have discrete spectrum .
its eigenvalues are given by @xmath58 , @xmath59 .
later one can take the limit @xmath60 .
consider the equations of motion for this model .
vacuum solutions ( with @xmath61 ) , are given by the commutative sets of operators @xmath62 . in terms of `` smooth '' functions and up to a gauge transformation ,
they are given by the arbitrary functions @xmath63 .
as it is generally known , two and more _ continuous _ functions of one variable always form a functional dependent set , therefore , the set of @xmath63 is a such one if @xmath63 are continuous functions of @xmath40 .
this property , however does not hold for the discontinuous functions .
after the kinetic term absorption the equations of motion are no more differential equations , thus the condition imposed one the solutions of equations of motion to be smooth functions are no more justified and can be given up .
moreover , the spectrum of @xmath40 itself is discrete . from the other hand , since the smoothness properties are strictly related to the notion of topology of the space - time , the non - smooth solution can be interpreted as the changing of the space - time topology .
the space - time topology can be extracted from the set of operators @xmath62 in the framework of connes approach @xcite .
consider that the solution @xmath62 now carries the attributes of the @xmath0 dimensional space , i.e. they form a @xmath0-dimensional lattice , and can be expressed as functions of the basic set of independent operators @xmath21 , @xmath13 . in what follows
let us construct an explicit solution with functionally independent ( and , therefore , discontinuous ) @xmath64 .
since eigenvalues of @xmath40 and @xmath21 form , respectively , one - dimensional and @xmath0 dimensional lattices , the solution @xmath64 is given explicitly by the map from the one - dimensional lattice @xmath65 of the eigenvalues of @xmath40 to the @xmath0-dimensional lattice @xmath66 of eigenvalues of @xmath21 . due to the reparameterisation invariance one can take both lattices to be regular and rectangular ones .
the map can be constructed through the following sequence of steps ( the particular case for @xmath67 is depicted in the figure ) . 1
. map the origin of @xmath65 to the origin of @xmath66 .
2 . map @xmath0 points next to origin in the positive direction and @xmath0 ones in negative direction of @xmath65 to the nearest neighbor points to the origin of @xmath66 , using e.g. lexicographic ordering .
3 . fill the the remaining points of the @xmath0-dimensional hypercube of two lattice unites size centered at origin of @xmath66 by images of @xmath68 points of @xmath65 .
4 . in the same manner
fill the next hyper - cubic shell of @xmath66 by images of of points of @xmath65 , etc .
thus , @xmath69 points around the origin of @xmath65 fill the hyper - cube of the size @xmath70 centered at the origin of @xmath66 .
as it can be seen by the construction the map is one - to - one .
however , the resulting operators @xmath21 are functionally independent . under this map small momenta
are transfered to small momenta , and large ones to large ones .
this means that respectively low / high energy states of one model are mapped into low / high energy sector of another one .
this correspondence allows one to pass from the one dimensional heisenberg algebra to the heisenberg algebra in arbitrary dimension @xmath54 ( or from two - dimensional noncommutative space to @xmath3 dimensional one ) . indeed , using the approach of ref .
@xcite one can introduce for each operator @xmath21 its canonical conjugate @xmath23 , satisfying the heisenberg commutation relation in @xmath0 dimensions , @xmath71=-{\mathrm{i}}\delta_i^j.\ ] ] one can , therefore , pass from ( [ s_op ] ) with @xmath55 to the equivalent description as a model in a different dimension @xmath54 .
let us shortly describe the construction @xcite . consider the eigenvalue problem for adjoint operators @xmath72 $ ] and @xmath73 $ ] .
this problem is consistent since @xmath74 and @xmath75 are commutative and self - adjoint on the space of square integrable operators with bounded @xmath74 .
the eigenvalue problem is solved by the ( eigen)operators , @xmath76 where @xmath77 and @xmath78 are eigenvalues of @xmath74 and @xmath79 respectively . since , by the construction , both @xmath41 and @xmath36
can be represented on the same hilbert space @xmath80 , ( in fact , we introduced isomorphism between @xmath80 and @xmath35 ) , one can expand an arbitrary square integrable operator @xmath81 of original model in the @xmath82 basis , using the trace over the hilbert space of the one - dimensional heisenberg algebra , and get an operator @xmath83 in the @xmath0-dimensional model , @xmath84 where , @xmath85 applying this procedure to the fields in model ( [ s_op ] ) with @xmath55 , one obtains the equivalent description in terms of a @xmath0 dimensional model with any @xmath54 .
formally , we have introduced here a noncommutative ( and discontinuous ) change of variables . indeed ,
since @xmath21 and @xmath23 are invertible functions of one - dimensional @xmath40 and @xmath17 , one can find their inverse , @xmath86 and @xmath87 , and plough this dependence in the one - dimensional operator @xmath81 to get the function @xmath88 which is a @xmath0 dimensional operator .
now we can give the following physical implication of this construction .
the @xmath0-dimensional noncommutative gauge model describes , in fact , the a d@xmath89-brane where @xmath90 @xcite .
we constructed a solution in the two - dimensional noncommutative gauge model which has the meaning of @xmath3 dimensional space .
this solution gives the correspondence between the gauge models ( gauge fields interacting with kaluza klein scalars ) in various even dimensions .
taking into account the brane interpretation of the noncommutative gauge models with scalar fields , this describes the condensation of d1-branes to a d@xmath89-brane , where @xmath91 . from this point of view the multiple vacua of the ikkt model ref .
@xcite , are nothing else than the condensation of d(-1)-branes described by the ikkt matrices to an arbitrary iib brane ( i.e. a brane with even - dimensional world - sheet ) , which is the @xmath3-dimensional reduction of the ten - dimensional noncommutative u(1 ) gauge model .
let us briefly summarise the results of this note .
first , we have shown that the model desribing the continuum limit of fluctuations of the ikkt matrix model around a commutative background , @xmath92=0 $ ] , in @xmath0 dimensions is equivalent to one describing fluctuations around a @xmath3 dimensional background satisfying , @xmath93={\mathrm{i}}b_{ij}$ ] , where @xmath94 is a nondegenerate antisymmetric scalar matrix .
`` physically '' this means that in the singular limit @xmath95 corresponds to doubling of the dimensionality of the noncommutative gauge model .
second , we demonstrated that the noncommutative model in @xmath96 dimensions can be isomorphically mapped to the @xmath3-dimensional one .
the last feature may be interpreted as the noncommutative geometry counterpart of the duality relating various branes or their condensation , @xcite . in early works on ikkt model @xcite ,
it was conjectured that this in the limit @xmath1 generates the space - time as the set of expectation values of operators @xmath6 . in this context
it seems natural that the topology and , in particular , the space - time dimensionality are also generated by the solution @xmath64 .
e. witten , ( 2000 ) , hep - th/0006071 , n. seiberg and e. witten , jhep 09 ( 1999 ) 032 , hep - th/9908142 , n. ishibashi et al .
b498 ( 1997 ) 467 , hep - th/9612115 , t. banks et al .
d55 ( 1997 ) 5112 , hep - th/9610043 , a. connes , m.r .
douglas and a. schwarz , jhep 02 ( 1998 ) 003 , hep - th/9711162 , j. ambjrn et al .
, jhep 05 ( 2000 ) 023 , hep - th/0004147 , c. sochichiu , jhep 05 ( 2000 ) 026 , hep - th/0004062 , c. sochichiu , phys .
b485 ( 2000 ) 202 , hep - th/0005156 , r. gopakumar , s. minwalla and a. strominger , jhep 05 ( 2000 ) 020 , hep - th/0003160 , a. connes , ( 2000 ) , hep - th/0003006 , l. alvarez - gaume and s.r .
wadia , ( 2000 ) , hep - th/0006219 , o.j .
ganor , s. ramgoolam and i. washington taylor , nucl .
b492 ( 1997 ) 191 , hep - th/9611202 , t. banks , n. seiberg and s. shenker , nucl .
b490 ( 1997 ) 91 , hep - th/9612157 , r. dijkgraaf , e. verlinde and h. verlinde , nucl .
62 ( 1998 ) 348 , hep - th/9709107 , | here we construct a map from the algebra of fields in two - dimensional noncommutative of u(1 ) yang mills fields interacting with kaluza klein scalars to a @xmath0-dimensional one , as a solution in the two - dimensional model .
this proves the equivalence of noncommutative models in various ( even ) dimensions .
physically this map describes condensation of d1-branes . |
the formation of the earliest galaxies plays a key role in a number of the most important questions being addressed in cosmology today .
the first galaxies are predicted to have been the dominant sources of the radiation which reionized the universe ( e.g. ciardi et al .
2006 ) , and they may have hosted the majority of primordial star formation ( greif & bromm 2006 ; but see also jimenez & haiman 2006 ) .
they are the likely sites for the formation of the most metal - poor stars that have recently been found in our galaxy ( e.g. christlieb et al .
2002 ; beers & christlieb 2005 ; frebel et al . 2005 ) , and possibly for the first efficient accretion onto the stellar black holes ( see johnson & bromm 2007 ) which may have been the seeds for the @xmath3 10@xmath6 m@xmath5 black holes that are inferred at redshifts @xmath1 @xmath0 6 ( fan et al .
2004 , 2006 ) .
furthermore , an understanding of the formation of the first galaxies is crucial for the interpretation of galaxies now beginning to be observed at @xmath7 6 ( e.g. mobasher et al .
2005 ; iye et al . 2006 ; bouwens & illingworth 2006 ) , as well as of the objects at redshifts @xmath7 10 which are expected to be detected with upcoming telescopes such as the _ james webb space telescope _ ( @xmath8 ) ( gardner et al .
2006 ) . among these systems
there promise to be some of the first metal - free objects that will be observable , and as such it is important that theoretical predictions of their properties are made . what were the effects of the radiative feedback from the first generations of stars on the formation of the first galaxies ?
it is now widely held that the first stars ( termed population iii ) were likely very massive , and therefore emitted copious amounts of radiation which profoundly affected their surroundings ( bromm et al . 1999 , 2002 ; abel et al . 2002 ; yoshida et al .
2006 ; gao et al .
recent work has demonstrated that the h ii regions surrounding the first stars were able to evacuate the primordial gas from the minihalos that hosted these objects ( whalen et al . 2004 ; kitayama et al .
2004 ; alvarez et al .
2006 ; abel et al .
the impact of these h ii regions on second generation star formation is complex ( e.g. ricotti , gnedin & shull 2001 ; oh & haiman 2003 ; ahn & shapiro 2006 ; susa & umemura 2006 ) .
while initially the density in these regions is suppressed and the gas within heated to @xmath9 k , vigorous molecule formation can take place once the gas begins to cool and recombine after the central pop iii star has collapsed to fom a massive black hole , leading to the possibility of the formation of low - mass primordial stars ( nagakura & omukai 2005 ; oshea et al .
2005 ; johnson & bromm 2006 , 2007 ; yoshida et al .
2007 ) .
an additional radiative feedback effect from the first stars is the photo - dissociation of the fragile hydrogen molecules which allow the primordial gas to cool and collapse into minihalos , with virial temperatures @xmath2 8,000 k ( e.g. barkana & loeb 2001 ) .
the effects of the molecule - dissociating radiation from the first stars can reach far beyond their h ii regions ( e.g. ciardi et al . 2000 ) , and thus star formation in distant minihalos may have been delayed or quenched altogether ( e.g. haiman et al .
1997 , 2000 ; mackey et al .
interestingly , however , while the general intergalactic medium ( igm ) at the epoch of the first stars becomes optically thick to lyman - werner ( lw ) photons only over vast distances ( e.g. haiman et al .
2000 ; see also glover & brand 2001 ) , the high molecule fraction that persists inside the first relic h ii regions leads to a high optical depth to these photons , potentially allowing star formation to take place in minihalos down to lower redshifts than would otherwise be possible ( ricotti et al .
2001 ; oh & haiman 2002 ; machacek et al .
2001 , 2003 ; johnson & bromm 2007 ) . in the present work , we self - consistently track the formation of , and the radiative feedback from , individual pop iii stars in the course of the formation of a primordial protogalaxy .
we compute in detail the h ii regions and lw bubbles of each of these sources , and follow the evolution of the primordial gas as it becomes incorporated into the protogalaxy . in 2 , we describe our numerical methodology .
our results are presented in 3 , while we summarize our conclusions and discuss their implications in 4 .
we employ the parallel version of gadget for our three - dimensional numerical simulations .
this code includes a tree , hierarchical gravity solver combined with the smoothed particle hydrodynamics ( sph ) method for tracking the evolution of the gas ( springel , yoshida & white 2001 ) . along with h@xmath10 , h@xmath11 ,
h , h@xmath12 , h@xmath13 , e@xmath12 , he , he@xmath14 , and he@xmath15 , we have included the five deuterium species d , d@xmath13 , d@xmath12 , hd and hd@xmath12 , using the same chemical network as in johnson & bromm ( 2006 , 2007 ) .
we carry out a three - dimensional cosmological simulation of high-@xmath1 structure formation which evolves both the dark matter and baryonic components , initialized according to the @xmath16cdm model at @xmath1 = 100 . as in earlier work
( bromm et al . 2003 ; johnson & bromm 2007 ) , we adopt the cosmological parameters @xmath17 , @xmath18 , @xmath19 , and @xmath20 , close to the values measured by _
wmap _ in its first year ( spergel et al . 2003 ) . here
we use a periodic box with a comoving size @xmath21 = 460 @xmath22 kpc , but unless stated explicitly , we will always refer to physical distances in the present work .
our simulation uses a number of particles @xmath23 = @xmath24 = 128@xmath25 , where the sph particle mass is @xmath26 @xmath3 740 @xmath27 .
we have determined the maximum density of gas that can be reliably resolved in this simulation by carrying out a cosmological simulation from @xmath1 = 100 , in which we allow the gas to cool and collapse into minihalos without including radiative effects .
we then compare the minimum resolved mass , which we take to be @xmath3 64 @xmath26 , with the bonnor - ebert mass , given by ( see , e.g. , palla 2002 ) @xmath28 where @xmath29 and @xmath30 are the number density and temperature of the gas , respectively .
as shown in figure 1 , the gas evolves according to the canonical behavior of primordial gas collapsing in minihalos ( see , e.g. , bromm et al .
we expect the gas in our simulations with radiative feedback to behave similarly as it collapses to high densities , since it is the formation of , and cooling by , molecules which will drive the collapse in both cases .
thus , we take the maximum density that we can reliably resolve to be that at which the bonnor - ebert mass becomes equal to the resolution mass . as is evident in figure 1
, this criterion results in a maximum resolvable density of @xmath31 @xmath3 20 @xmath32 .
this density is four orders of magnitude higher than the background mean density at a redshift of @xmath1 @xmath0 15 , and such overdensities only occur in the minihalos within which the first pop iii stars form ( see , e.g. , bromm & larson 2004 ; yoshida et al .
we take it here that one pop iii star , assumed to have a mass of 100 m@xmath5 , will form from this dense , collapsing primordial gas inside a minihalo , consistent with recent work which shows that , in general , only single stars are expected to form in minihalos under these conditions ( yoshida et al .
pop iii stars with this mass are predicted to directly collapse to a black hole , and therefore produce no supernova explosion ( e.g. heger et al .
2003 ) , which allows us to self - consistently neglect the possibility of ejection of metals into the primordial gas .
we will consider this possibility in future work . in the present work
, we focus on the radiative feedback from the first stars . in our simulations with radiative feedback
, we assume that stars are formed in minihalos which acquire densities higher than @xmath31 = 20 @xmath32 . in order to account for the radiative feedback from a star formed in a minihalo ,
the gas surrounding the star is first photo - heated .
we then calculate the extent of the h ii region , as well as of the lw bubble around the star .
we carry out this procedure every time a star forms in the simulation .
the pop iii star will soon die , and we then let the simulation evolve once more , allowing recombination to take place in the relic h ii region , and for molecules to reform within the relic lw bubble .
we expect that this procedure will provide reliable results , as the @xmath33 myr lifetime of a pop iii star is short compared to the typical dynamical times of the gas in this simulation . to account for the presence of a 100 m@xmath5 pop iii star in the minihalos in which the gas collapses to a density of @xmath31 , we first photoheat and photoionize the gas within 500 pc of the gas particle which first reaches this density for a duration of the lifetime of the star , using the same heating and ionization rates as in johnson & bromm ( 2007 ) .
our choice of a 500 pc radius ensures that the entire gas within the source minihalo , with virial radius @xmath34 pc , is photoheated , but that we do not photoheat the dense , neutral gas in neighboring halos . just as in this previous work , we reproduce the basic density and velocity structure of the gas within 500 pc of the central source that has been found in detailed one - dimensional radiation hydrodynamics calculations ( kitayama et al .
2004 ; whalen et al . 2004 ) .
once this density structure is in place around the point source , we employ a ray - tracing technique to solve for the h ii region that surrounds the star at the end of its life .
we cast rays in @xmath35 @xmath3 100,000 directions from the central source , and divide each ray into 500 segments .
then , we add up all of the recombinations that take place over the course of the star s lifetime in each bin along each of the @xmath35 rays , taking the number of recombinations to be @xmath36 where @xmath37 is the case b recombination coefficient , @xmath38 is the distance of the bin from the star , @xmath39 is length of the bin in the direction radial to the star , and @xmath40 is the lifetime of the star , here taken to be 3 myr ( schaerer 2002 ) .
we compute @xmath41 , the average number density of hydrogen atoms in a bin , as @xmath42 here @xmath43 is the number of sph particles in the bin and @xmath44 is the number density of hydrogen of the individual sph particles in that bin .
next , we assume that the star radiates an equal number of photons in every direction , and we take the total number of ionizing photons that it radiates in its lifetime to be @xmath45 where we have chosen @xmath46 , the average number of ionizing photons emitted per second by the star , to be 10@xmath47 s@xmath48 ( see bromm et al .
2001 ; schaerer 2002 ) .
we then add up the recombinations in all of the bins , along each of the rays , beginning with those closest to the star and moving outward , until the number of recombinations along a ray equals the number of ionizing photons that are emitted along that ray .
if the number of recombinations in the bin falls below the number of atoms in the bin , then we count the number of atoms in the bin against the number of photons as well . doing this for each of the rays , we solve in detail for the h ii region of the star .
we set the free electron fraction to unity for each of the sph particles that lie within the h ii region .
we set the temperature of the sph particles within the h ii region , but outside of the 500 pc photo - heated region , to @xmath30 = 18,000 k , roughly the value at the outer edge of the photo - heated region .
as well , the fraction of molecules in the h ii region is set to zero , as we assume that all molecules are collisionally dissociated at the high temperatures in the h ii region . to find the region in which the lw radiation from the star destroys h@xmath10 and hd molecules , the `` lw bubble '' in our terminology , we carry out a ray - tracing procedure similar to the one used to solve for the h ii region
we use the same bins as in that procedure , but now we evaluate the formation time of h@xmath49 molecules in each bin and compare this both to the lifetime of the star and to the dissociation time of the molecules . for each bin , we compute the h@xmath49 formation time as @xmath50 where @xmath51 and @xmath52 are the number densities of h@xmath11 and h@xmath53 , respectively . the sum is over all the particles in the bin , @xmath43 , and @xmath54 and @xmath55 are the rate coefficients for the following two main reactions that produce h@xmath10 : @xmath56 @xmath57 we adopt the following values for these rate coefficients ( de jong 1972 ; karpas et al . 1979 ; haiman et al .
1996 ) : @xmath58 @xmath59 the dissociation time for the molecules is obtained by finding the flux of lw photons from a 100 m@xmath5 pop iii star , assumed to be a blackbody emitter with radius @xmath60 and effective temperature @xmath61 k ( e.g. , bromm et al .
the dissociation time for unshielded molecules at a distance @xmath62 from the star is then given by ( abel et al .
1997 ) @xmath63 next , we note that for molecules to be effectively dissociated by the lw radiation , the dissociation time of the molecules must be shorter than both the lifetime of the star and the formation time of the molecules .
therefore , we compare all of these timescales for each bin along each ray and set the fraction of molecules to zero if @xmath64 @xmath2 @xmath65 and @xmath64 @xmath2 @xmath40 .
if this condition is not satisfied , then the molecule fraction is left unchanged from its value before the formation of the star .
this allows for the possibility of the effective shielding of h@xmath10 molecules because it accounts for the build - up of h@xmath10 column density , for instance , in relic h ii regions or in collapsing minihalos where the formation time of h@xmath10 is relatively short .
we take into account the effects of self - shielding by adding up the h@xmath49 column density @xmath66(h@xmath49 ) along the ray contributed by each bin in which the molecules are not effectively dissociated .
we then adjust the dissociation time for the molecules in shielded bins according to ( draine & bertoldi 1996 ) : @xmath67 when the column density of molecules between the bin and the star is @xmath66(h@xmath49 ) @xmath0 10@xmath68 @xmath69 .
the ionic species h@xmath53 and h@xmath11 , which are reactants in the main reactions which form h@xmath10 , can also , in principle , be destroyed by the radiation from the star .
the photo - dissociation times for these species are given in terms of the temperature of the star @xmath70 , the source of thermal radiation in our case , and the distance from the star @xmath62 , as ( dunn 1968 ; de jong 1972 ; galli & palla 1998 ) @xmath71 @xmath72 for the @xmath73 star , we find that @xmath74 @xmath3 5 @xmath75 10@xmath25 yr @xmath76 and @xmath77 @xmath3 9 @xmath75 10@xmath78 yr @xmath79 .
the formation times for these species , on the other hand , are @xmath80 @xmath81 for primordial gas at a temperature of @xmath30 = 100 k and a density of @xmath44 = 10@xmath82 @xmath32 , typical for gas at the outskirts of a collapsing minihalo .
these formation timescales become much shorter for gas deeper inside minihalos , where the densities and temperatures are generally higher .
it is in these regions , in and around minihalos , where the presence of molecules is most important for cooling the gas . within these regions
the photo - dissociation times for these ionic species are less than their formation times only if they are located @xmath2 2 kpc from the star .
thus , photo - dissociation of these species will become ineffective at distances @xmath0 2 kpc from the star , a distance comparable to the size of the h ii region of a pop iii star ( alvarez et al .
2006 ; abel et al . 2006 ) .
since we assume that molecules are collisionally destroyed inside the h ii region , and since the lw bubble will generally be larger than the h ii region , we ignore the photo - dissociation of h@xmath53 and h@xmath11 in our calculations .
lw photons can also be absorbed by hydrogen atoms , through the lyman series transitions , as discussed in detail by haiman et al .
( 1997 , 2000 ) .
however , this atomic absorption will only have a significant effect on the lw flux over distances large enough that the hubble expansion causes many of the lw photons to redshift to wavelengths of the lyman series transitions .
the light - crossing time for our cosmological box is much shorter than the hubble time at the redshifts that we consider .
thus , lw photons will be negligibly redshifted as they cross our cosmological box and we can safely neglect the minimal atomic absorption of these photons that may take place .
it has also been found that a shell of h@xmath10 molecules may form ahead of the expanding h ii regions surrounding the first stars ( see ricotti et al .
these authors find that such shells may become optically thick to lw photons .
however , kitayama et al . ( 2004 ) have discussed that such shells are likely short - lived , persisting for only a small fraction of the lifetime of the star .
thus , for our calculations we neglect the possible formation of such a shell , as we expect that the opacity to lw photons through this shell will be very small when averaged over the lifetime of the star .
additionally , as we show in section 3.1 , the regions affected by the lw feedback from a single pop iii star extend , at most , only a few kiloparsecs beyond the h ii region of such a star , which itself extends @xmath3 5 kpc .
if an h@xmath10 shell forms ahead of the h ii region , then the extent of the lw bubble will only be suppressed by , at most , a factor of a few in radius .
we have carried out two simulations , one with radiative feedback and one in which the simulation evolves without including radiative effects .
for the former simulation , we allow stars to form when the density reaches @xmath31 , and the expansion of the gas around the star due to photo - heating suppresses the density so that our resolution limit is not violated . for the simulation without radiative feedback
we allow sink particles to form when the density of the gas reaches @xmath31 . since the sink particles will form only in minihalos which are expected to form pop iii stars , we are able to track the star formation rate in the case without feedback by tracking the formation of sink particles
. we can then compare the sites , and rates , of star formation in each of the simulations in order to elucidate the effect that radiative feedback has on pop iii star formation .
in this section we discuss the evolution of the primordial gas under the influence of the radiative feedback which arises as the first stars are formed in a region destined to subsequently be incorporated into the first protogalaxy .
indeed , other effects will become important in the course of the buildup of the first galaxies . among them
is the ejection of metals into these systems by the first supernovae ( e.g. bromm et al . 2003 ) .
however , we consider the early regime in which pop iii star formation dominates , and the effects of metals might not yet be important .
initially only taking into account the stellar radiative feedback , and neglecting chemical enrichment , relies on our simplifying assumption that only 100 m@xmath5 black hole - forming pop iii stars form , which are predicted not to yield supernovae , and therefore not to eject metals into their surroundings ( heger et al .
2003 ) .
although the initial mass function ( imf ) of the first generation of stars is not known with any certainty yet , there is mounting theoretical evidence that pop iii stars were very massive , and thus it is very likely that many of these stars ended there lives by collapsing directly to black holes , emitting few or no metals into the igm ( see fryer et al .
2001 ; heger et al .
2003 ) . here
we assume that all of the stars that form within our cosmological box are black hole - forming stars which do not enrich the igm with metals , and which therefore allow subsequent metal - free star formation to occur .
eventually , however , stars which create supernovae will form , and the ejected metals will be incorporated into the first protogalaxies , thus drawing the epoch of metal - free star formation to a close . in light of this , we end our simulation after the formation of the eighth star in our box at a redshift of @xmath1 @xmath3 18 , as we expect that at lower redshifts the effects of the first metals ejected into the primordial gas will become important ( but see jappsen et al .
also , at lower redshifts global lw feedback , due to star formation at distances far larger than our cosmological box , will become increasingly important .
that said , by tracking the formation of individual pop iii stars in our box , we are able to find a variety of novel results concerning the local radiative feedback from the first generations of stars
. the first star appears in our cosmological box at a redshift of @xmath1 @xmath3 23 .
it forms inside a minihalo with a total mass @xmath2 10@xmath4 m@xmath5 and the gas within this halo is evaporated due to the photo - heating from the star .
the h ii region that is formed around the star can be seen in figure 2 , which shows the electron fraction , h@xmath10 fraction , temperature , and density of the gas , in projection .
the h ii region , which has a morphology similar to those found in previous studies , extends out to @xmath3 4 kpc from the star , also similar to results found in previous works ( alvarez et al .
2006 ; wise et al . 2006
; yoshida et al . 2007 ) . as shown in figure 2 ,
the molecules within @xmath3 5 kpc are photodissociated by the lw radiation from the first star , and the lw bubble extends to only @xmath3 1 kpc outside of the h ii region . noting that the formation timescale for h@xmath10 in the neutral igm at these redshifts is of the order of @xmath3 300 myr , much longer than the lifetime of the massive stars that we consider here
, we can estimate the distance through the igm that the lw bubble should extend , @xmath83 , by evaluating the criterion for the effective dissociation of molecules at this distance from the first star : @xmath64 = @xmath40 . using equation ( 8) and taking the lifetime of the star to be 3 myr gives @xmath83
@xmath3 5 kpc , consistent with the result we find for the first star , shown in figure 2 ( see also ferrara 1998 ) . outside of this lw bubble molecules will not be dissociated effectively by the single pop iii star , owing largely to its short lifetime . only when continuous star formation sets in will the lw bubbles of the first generations of stars merge and become large enough to establish a more pervasive lw background flux ( see e.g. haiman et al .
2000 ) .
the properties of the primordial gas within our box are strongly time - dependent , as any gradual evolution of the gas is disrupted each time a star turns on and heats the gas , ionizes atoms , and photodissociates molecules .
certain robust patterns , however , do emerge in the course of the evolution of the primordial gas .
figure 3 shows the chemical and thermal properties of the gas at a redshift @xmath1 @xmath3 18 , just after the death of the eighth star in our cosmological box . here , the light - shaded particles are those which have been contained within an h ii region , and so have passed through a fully ionized phase . the ionized gas in the h ii regions begins to recombine and cool once the central star dies .
the dynamical expansion of these hot regions leads to the adiabatic cooling of the gas , as can be seen in the upper left panel of figure 3 .
the plot shows relic h ii regions at different evolutionary stages .
the older ones are generally cooler , owing to the molecular cooling that has had more time to lower the temperature of the gas . indeed , the first relic h ii regions by this redshift , @xmath84 myr after the first star formed , have already cooled to near the temperature of the un - ionized gas .
the electron fraction of the relic h ii region gas , however , is still much higher than that of the un - ionized gas , as can be seen in the upper - right panel of figure 3 . that the cooling of the gas occurs faster than its recombination leads to the rapid formation of molecules ( e.g. kang & shapiro 1992 ; oh & haiman 2003 ; nagakura & omukai 2005 ; johnson & bromm 2006 ) .
this elevated fraction of both h@xmath10 and hd molecules in the relic h ii region gas is evident in the bottom panels of figure 3 . the high abundance of molecules in relic h ii regions can lead to efficient cooling of the gas , and this has important consequences in the first protogalaxies .
in particular , a high fraction of hd in these regions could allow the gas to cool to the temperature of the cosmic microwave background ( cmb ) , @xmath85 , the lowest temperature attainable by radiative cooling , and this effective cooling may lead to the formation of lower mass metal - free stars ( nagakura & omukai 2005 ; johnson & bromm 2006 ; yoshida 2006 ) .
indeed , figure 3 shows that the hd fraction can greatly exceed the minimum value needed for efficient cooling to the cmb temperature floor in local thermodynamic equilibrium ( lte ) , @xmath86 @xmath3 10@xmath87 ( johnson & bromm 2006 ) . while the lw feedback from the stars that form in our box can very effectively destroy molecules within @xmath3 5 kpc of the stars by the end of their lives , this feedback is not continuous . following the death of a given star , the molecules will begin to reform in the absence of lw radiation .
the time required for the formation of h@xmath10 molecules is sensitively dependent on the ionized fraction of the gas , but the formation time can be relatively short for un - ionized gas at high densities , as well . in relic h ii regions the fraction of h@xmath10 can reach 10@xmath88 within @xmath3 1 myr ( johnson & bromm 2007 ) . in collapsing minihalos , where the molecules play a key role in cooling the gas and allowing it to continue collapsing , the formation times are in general longer at the densities we consider here , @xmath29 @xmath2 20 @xmath32 .
we find that the formation timescale for un - ionized gas collapsing in minihalos is @xmath65 @xmath3 5 @xmath75 10@xmath89 yr at a density of 1 @xmath32 and a temperature of 900 k , and @xmath65 @xmath3 7 @xmath75 10@xmath4 yr at a density of 0.1 @xmath32 and a temperature of 500 k. the average time between the formation of stars in our box is @xmath90 myr , and so the molecules inside sufficiently dense minihalos can often reform and allow the gas to continue cooling and collapsing , in spite of the intermittent lw feedback from local star forming regions . in order to evaluate the possible effects of continuous lw feedback from sources outside of our box , we have carried out simulations
in which we include a lw background which destroys h@xmath10 molecules at a rate given by ( abel et al .
1997 ) @xmath91 where @xmath92 is the flux of lw photons in units of 10@xmath93 ergs s@xmath48 @xmath69 hz@xmath48 sr@xmath48 .
we have carried out simulations in which the value of @xmath92 is taken to be zero before the formation of the first star and 0.1 , 10@xmath82 , and 10@xmath94 afterwards , when a lw background might be expected to begin building up due to distant star formation .
for each of these simulations , we found the formation redshift of the second star in the box to be @xmath95 = 16.3 , 20 , and 20.5 , respectively . in our main simulation ,
in which we neglect a possible background lw flux , the second star formed at @xmath95 = 20.6 .
this demonstrates that a background lw flux of @xmath92 @xmath2 10@xmath82 would likely have little impact on our results , while a larger lw flux would simply delay the collapse of gas into minihalos and so lower the overall star formation rate in our box , consistent with previous findings ( see e.g. machacek et al .
2001 ; mesinger et al .
we emphasize , however , that for a substantial lw background to be established , a relatively high continuous star formation rate must be achieved , as we have shown that individual pop iii stars can only be expected to destroy molecules within @xmath3 5 kpc of their formation sites . in the very early stages of the first star formation , when short - lived single stars are forming in individual minihalos
( see e.g. yoshida et al .
2006 ) , it appears unlikely that a substantial lw background would be established , the feedback from the sources being instead largely local .
it may be only later , when continuous star formation begins to occur in larger mass systems that a pervasive lw background would likely be built up ( see , e.g. , haiman et al .
2000 ; greif & bromm 2006 ) .
as star formation continues , the volume occupied by relic h ii regions increases . because of the high molecule fraction that can develop in these regions , owing to the large electron fraction that persists for @xmath2 500 myr ( johnson & bromm 2007 ) , the increasing volume of the igm occupied by relic h ii regions implies an increase in the optical depth to lw photons in the vicinity of the first star formation sites . by a redshift of @xmath1
@xmath3 18 , eight stars have formed in our cosmological box and each has left behind a relic h ii region . as can be seen in figures 3 and 4
, the gas inside the relic h ii regions that have formed contains an h@xmath10 fraction generally higher than the primordial abundance of 10@xmath96 , and up to an abundance of @xmath3 10@xmath94 in the denser regions .
this elevated fraction of h@xmath10 inside the relic h ii regions leads to a high optical depth to lw photons , @xmath97 , through the relic h ii regions .
the column density through a relic h ii region which recombines in the absence of lw radiation can become of the order of @xmath98 10@xmath99 @xmath69 ( johnson & bromm 2007 ) .
because the molecules in the relic h ii regions that we consider here are subject to lw feedback from neighboring star formation regions , the optical depth through these regions may in general be lower .
however , the rapid rate of molecule formation in these regions , even considering the lw feedback from local star forming regions in our box , allows the molecule fraction to approach 10@xmath88 as late as @xmath3 100 myr after the death of the central star . this elevated molecule fraction combined with the growing volume - filling fraction of relic h ii regions leads to an appreciable optical depth to lw photons , which generally increases with time as more stars form and
create more relic h ii regions . to quantify this effect
, we calculate the average column density of h@xmath10 molecules through a cubic region of side length @xmath100 as the product of the length @xmath100 and the volume averaged number density of h@xmath10 molecules , given by @xmath101 where the sum is over all of the sph particles in the volume and @xmath102 is the number density of h@xmath10 at each of the sph particles . the volume associated with each individual sph particle , @xmath103 ,
is estimated as @xmath103 @xmath104 m@xmath105/@xmath106 , where m@xmath105 is the mass of the sph particle and @xmath106 is the mass density of the gas at that particle . the optical depth to lw photons
is then computed as ( draine & bertoldi 1996 ; haiman et al .
2000 ) @xmath107 figure 5 shows the optical depth to lw photons avergaged both over the central comoving 153 kpc @xmath22 of our cosmological box , in which the first star forms , and over the entire box , for which the comoving side length is 460 kpc @xmath22 . before the formation of the first star
, the optical depth evolves largely owing to the cosmic expansion , following the relation @xmath97
@xmath104 @xmath102 @xmath100 @xmath108 ( 1+@xmath1)@xmath78 , because the average h@xmath10 fraction does not change appreciably . however , with the formation of the first star in our box at @xmath1 @xmath104 23 the optical depth begins to change dramatically in the inner portion of the box , first falling to a value of @xmath104 0.1 due to the lw feedback from the first star and then steadily climbing to values @xmath0 2 as copious amounts of molecules form inside the relic h ii regions that accumulate as star formation continues .
the evolution of the optical depth averaged over the entire box is not as dramatic , as the fraction of the volume of the whole box occupied by relic h ii regions is much smaller than the fraction of the central region that is occupied by these molecule - rich regions .
however , the optical depth averaged over the whole box , which is a better estimate of the optical depth over cosmological distances , still rises to @xmath97 @xmath0 1.5 across our box , an appreciable value which will serve to impede the build - up of a cosmologically pervasive lw background .
accretion onto pop iii relic black holes may be inefficient for some time following the formation of these objects , owing to the fact that pop iii stars photo - heat and evaporate the gas within the minihalos which host them ( johnson & bromm 2007 ; see also yoshida 2006 ) . indeed , accretion onto pop iii relic black holes at close to the eddington limit
can only occur if the accreted gas has a density above @xmath3 10@xmath78 @xmath32 , and it is only in collapsing halos that such densities are achieved at the high redshifts at which the first stars formed ( johnson & bromm 2007 ) . by assumption ,
all of the stars that are formed in our simulation are black hole - forming pop iii stars .
if these black holes remain inside their host minihalos , then by tracking the evolution of the gas within these photo - evaporated host minihalos , we can learn when efficient accretion onto these pop iii relic black holes may occur .
the minihalo within which the first star forms at a redshift @xmath1 @xmath3 23 resides within the relic h ii region left by the first star . due to the formation of a high fraction of molecules , and to the molecular cooling that ensues , the relic h ii region gas cools down to temperatures @xmath3 10@xmath25 k , below the virial temperature of this 10@xmath4 m@xmath5 minihalo .
the gas then re - collapses into the minihalo , reaching a peak density of @xmath31 at a redshift @xmath1 @xmath3 19 , or @xmath3 50 myr after the formation of the first star .
figure 6 shows the properties of the relic h ii region gas as a function of distance from the center of this minihalo , at the time when the gas has collapsed to a density of @xmath31 .
we can not , with this simulation , resolve what happens once the gas collapses further and reaches higher densities .
however , we can estimate the time it will take for the gas at the center of the halo to reach a density of @xmath29 @xmath3 10@xmath78 @xmath32 as the free fall time of the gas , which is @xmath109 @xmath3 10 myr .
thus , a pop iii relic black hole at the center of this halo could be expected to begin accreting gas efficiently @xmath3 60 myr after its formation .
this is a significant delay , and could pose serious challenges to theories which predict that efficient accretion onto pop iii relic black holes can lead to these black holes becoming the supermassive black holes that power the quasars observed in the _ sloan digital sky survey _ at redshifts @xmath1 @xmath0 6 ( e.g. yoshida 2006 ; johnson & bromm 2007 ; li et al . 2006 ) .
while abundant molecules can form within relic h ii regions , the lw feedback from neighboring star - forming regions can suppress the effect of this elevated fraction of molecules .
the electron fraction remains high in relic h ii regions for up to @xmath3 500 myr in the general igm , but in higher density regions where the gas is recollapsing , the electron fraction drops much more quickly .
figures 3 and 6 show that the electron fraction drops to a value of @xmath2 10@xmath88 , comparable to the electron fraction of the un - ionized gas , once the density of the relic h ii region gas becomes @xmath0 10 @xmath32 .
thus , once the gas reaches these densities the ionized fraction will become too low to catalyze the formation of a high fraction of molecules , and of hd molecules in particular .
therefore , in order for hd to be an effective coolant of the primordial gas in relic h ii regions , the abundanct hd molecules that are formed at densities @xmath2 10 @xmath32 must not be destroyed by lw feedback from neighboring star - forming regions before the gas collapses to high densities and forms stars .
if we estimate the timescale on which the relic h ii region gas would collapse to form stars as the free - fall time of the gas , we find that the molecules must be shielded from photodissociating radiation for at least @xmath109 @xmath0 10 myr in order for the high abundance of hd molecules to persist , so that the formation of so - called pop ii.5 stars might be enabled , with their hypothesized masses of @xmath110 ( see johnson & bromm 2006 ) .
we find that the relic h ii region gas that re - collapses into the minihalo in which the first star formed , shown in figure 6 , carries a high fraction of hd molecules , as lw feedback from neighboring stars does not effectively dissociate the molecules in this relatively dense and self - shielded gas .
the hd fraction exceeds 10@xmath111 , becoming an order of magnitude higher than its value for un - ionized primordial gas collapsing in a minihalo .
thus , for this case , hd cooling will likely be effective at higher densities as the gas collapses further , and we expect that a pop ii.5 star , with a mass of the order of 10 m@xmath5 , might form later on , if we were to run the simulation further ( nagakura & omukai 2005 ; yoshida 2006 ; johnson & bromm 2006 ; yoshida et al .
had star formation taken place nearer this minihalo between the formation of the first star and the re - collapse of the gas into the host minihalo , then the molecule fraction would likely not be so elevated , and a higher mass metal - free star would be more likely to form . thus , while in our simulation it appears that the first relic h ii region that forms may give rise to pop ii.5 star formation , we emphasize that the possibility of the formation of pop ii.5 stars in relic h ii regions is very dependent on the specific lw feedback that affects the gas in these regions . to discern the effect the local radiative feedback from the first stars has on the star formation rate , we have compared the results obtained from our simulations with and without radiative feedback . by a redshift of @xmath1
@xmath3 18 , a total of nine star - forming regions were identified in our simulation without feedback , while at the same epoch eight stars had formed in our simulation including feedback .
thus , we find that the average star formation rate at redshifts @xmath1 @xmath0 18 is diminished by a factor of perhaps only @xmath2 20 percent due to local radiative feedback , although this result is subject to the small number statistics within our single cosmological box .
figure 7 shows the locations of the sites of star formation for both cases , plotted in comoving coordinates against the projected density field .
the orange squares denote sites where pop iii stars could have formed in the case without feedback , while the green dots denote sites where pop iii stars formed in the simulation including radiative feedback .
thus , the sites where star formation is suppressed by the radiative feedback are marked by the orange squares which are not filled by a green dot .
we point out , however , that we do not include lw feedback from stars which may have formed outside of our box , and hence it is possible that the overall lw feedback may be stronger than we find here . at redshift
@xmath1 @xmath3 18 , we end the simulation , but note that star formation will likely take place at an increasing rate as the collapse fraction increases with time .
this could lead to a continuous lw background produced within our box , different from the intermittent lw feedback produced by individual stars that occurs in the simulation down to @xmath1 @xmath3 18 .
also , the limits of our resolution prohibit us from discerning the stronger shielding of h@xmath10 molecules and i - front trapping that could occur within very dense collapsing minihalos ( see ahn & shapiro 2006 ) .
however , we note that these authors find that the radiative feedback on collapsing minihalos from nearby stars generally does not greatly affect the final outcome of the collapse , as halos which collapse in the absence of radiative feedback generally also collapse when radiative feedback is applied ( see also susa & umemura 2006 ) , roughly consistent with our results in the present work . thus , our limited resolution may not substantially impact the results that we find for the slight suppression of star formation due to local radiative feedback , although higher resolution simulations will be necessary to more precisely study the full impact of radiative feedback from the first stars on the first protogalaxies .
we have performed cosmological simulations which self - consistently account for the radiative feedback from individual pop iii stars , as they form in the course of the assembly of the first protogalaxies .
we have solved in detail for the h ii regions , as well as for the lw bubbles of these stars , wherein molecule - dissociating radiation effectively destroys h@xmath10 and hd molecules .
the local radiative feedback from the first stars is complex , and we find a variety of novel results on the evolution of the primordial gas , on the effects of the lw radiation from the first stars , on the nature of second generation star formation , and on black hole accretion . while the lw radiation from the first stars can , in principle , greatly suppress pop iii star formation in the early universe , we find that a number of factors minimize the effectiveness of this negative feedback .
firstly , the lw radiation produced locally by individual stars is not uniform and constant , as lw feedback has been modeled in previous work ( e.g. , ciardi & ferrara 2005 ; mesinger et al .
2006 ) , but rather is present only during the brief lifetimes of the individual stars that produce it .
thus , even if the molecules in collapsing minihalos and relic h ii regions are destroyed by the radiation from individual stars , they will , at the early stages of pop iii star formation , have time to reform and continue cooling the primordial gas in between the times of formation of local stars . furthermore , because the lw bubbles of individual pop iii stars extend only to @xmath83 @xmath3 5 kpc from these sources , due to the short stellar lifetimes , the build up of a pervasive lw background would likely have to await the epoch of continuous star formation , which is fundamentally different from the epoch of the first stars , in which these sources shine for short periods within individual minihalos . as star formation continues , the volume - filling fraction of relic h ii regions increases as well , and this , combined with the high fraction of molecules that form in these regions , leads to an opacity to lw photons through the igm which increases with time .
this opacity can become of the order of @xmath97 @xmath0 2 through individual relic h ii regions ( johnson & bromm 2007 ; see also ricotti et al .
2001 ; machacek et al .
2001 , 2003 ; oh & haiman 2002 ) . furthermore , as the volume - filling fraction of relic h ii regions increases with time , @xmath97 through the general igm may become similarly large , and this effect will have to be considered in future work which seeks to elucidate the effect of lw feedback on pop iii star formation .
we find that metal - free stars with masses of the order of 10 m@xmath5 , the postulated pop ii.5 stars ( e.g. johnson & bromm 2006 ; greif & bromm 2006 ) , might form from the molecule - enriched gas within the first relic h ii regions , although we note that this may not occur in general , due to lw feedback from neighboring star - forming regions .
this susceptibility to lw feedback is due to the fact that the high fraction of hd molecules which forms in the electron - rich , low density regions of relic h ii regions must persist until this gas has had time to collapse to high densities and form stars .
if lw feedback from neighboring star - forming regions destroys the molecules after the gas has collapsed to densities @xmath0 10 @xmath32 , then the abundance of hd molecules will not likely be elevated when the gas forms stars and pop ii.5 star formation may be suppressed . because this implies that the molecules must be shielded from lw radiation for , at least , the free - fall time for gas with densities @xmath2 10 @xmath32 , or @xmath0 10 myr , we conclude that pop ii.5 star formation in relic h ii regions may occur only in the circumstances when local pop iii star formation is suppressed over such timescales .
however , we also point out that the shielding provided by the high h@xmath10 fraction in these relic h ii regions may help to minimize the lw feedback from neighboring star - forming regions , and so may make pop ii.5 star formation in relic h ii regions possible in many cases .
we find that the ionized primordial gas surrounding the first star formed in our simulation , at a redshift of @xmath1 @xmath3 23 , recombines and cools by molecular cooling to temperatures below the virial temperature of the minihalo that hosted this first star .
thus , this relic h ii region gas is able to re - collapse to densities @xmath0 20 @xmath32 within this minihalo after @xmath3 50 myr from the death of the star .
it is predicted that many pop iii stars will collapse directly to form black holes with masses of the order of 100 m@xmath5 ( e.g. heger et al .
2003 ) , and if such a black hole resides within this host halo , then we find that it may begin accreting dense primordial gas at close to the eddington rate after @xmath3 60 myr from the time of its formation .
this is an important consideration to be incorporated into models of the growth of the @xmath112 black holes which have been observed at redshifts @xmath113 , as it places constraints on the amount of matter that a given relic pop iii black hole could accrete by this redshift ( e.g. haiman & loeb 2001 ; volonteri & rees 2006 ; li et al .
2006 ) . finally , by comparing the star formation rates which we derive from our simulation including radiative feedback with those derived from our simulation in which feedback is left out , we have seen that local radiative feedback from the first stars likely only diminishes the pop iii star formation rate by a factor of , at most , a few . in our simulation , in particular , we find that this rate is decreased by only @xmath2 20 percent , although this may be less suppression than would be expected by the overall radiative feedback , as we did not include the possible effects of a global lw background .
future simulations which resolve densities higher than those reached here , and which self - consistently track the build - up of the lw background along with the igm opacity to lw radiation , will be necessary to more fully explore the radiative effects of the first stars on the formation of the first galaxies .
however , the goal of understanding the formation of the first galaxies is now clearly getting within reach , and the pace of progress is expected to be rapid . | the formation of the first galaxies is influenced by the radiative feedback from the first generations of stars .
this feedback is manifested by the heating and ionization of the gas which lies within the h ii regions surrounding the first stars , as well as by the photodissociation of hydrogen molecules within the larger lyman - werner ( lw ) bubbles that surround these sources . using a ray - tracing method in three - dimensional cosmological simulations , we self - consistently track the formation of , and radiative feedback from , individual stars in the course of the formation of a protogalaxy .
we compute in detail the h ii regions of each of these sources , as well as the regions affected by their molecule - dissociating radiation .
we follow the thermal , chemical , and dynamical evolution of the primordial gas as it becomes incorporated into the protogalaxy .
while the igm is , in general , optically thick to lw photons only over physical distances of @xmath0 30 kpc at redshifts @xmath1 @xmath2 20 , the high molecule fraction that is built up in relic h ii regions and their increasing volume - filling factor renders even the local igm optically thick to lw photons over physical distances of the order of a few kiloparsecs .
we find that efficient accretion onto pop iii relic black holes may occur after @xmath3 60 myr from the time of their formation , by which time the photo - heated relic h ii region gas can cool and re - collapse into the 10@xmath4 m@xmath5 minihalo which hosts the black hole .
also , pop ii.5 stars , postulated to have masses of the order of 10 m@xmath5 , can likely form from this re - collapsing relic h ii region gas , but their formation may be suppressed by lw feedback from neighboring star - forming regions .
overall , we find that the local radiative feedback from the first generations of stars suppresses the star formation rate by only a factor of , at most , a few . |
bose - einstein condensates loaded into optical lattices , which perfectly realize spatially periodic potentials , represent an exciting field of research in the sense that many simplified toy models of condensed matter physics can be studied in an exceptionally clean manner @xcite .
this is achieved by modern means of atom and quantum optics that allow the experimentalist an unprecedented control of initial conditions in coordinate and momentum space and also of the desired dynamics @xcite .
a paradigm of quantum transport on the microscopic scale is the wannier - stark ( ws ) problem , where particles move in a tilted , but otherwise spatially periodic potential .
the famous bloch oscillations and related phenomena , such as interband tunneling , were observed in experiments with quasi - particles in superlattices @xcite , with light in optical nonlinear media @xcite , and in great detail also with ultracold atoms moving in optical lattices @xcite .
all experimental studies based on the latter realization were performed in a regime where atom - atom interactions are either negligible @xcite or they reduce to a perturbative mean - field effect @xcite . the regime of strong correlations in the ws system , in which interactions can not be reduced to a mean - field model or even dominate the evolution has been addressed only theoretically up to now @xcite .
state - of - the - art experiments are , however , capable of getting into a regime of filling factors ( i.e. of atoms per lattice site ) of the order one , where interaction - induced correlations are crucial @xcite . motivated by the experimental progress , we extend previous studies of the asymmetric triple - well @xcite and of the ws problem @xcite . in the present work we give a more comprehensive and quantitative account of our findings briefly reported in @xcite . in section
[ sec : oneband ] we introduce our two - bands bose - hubbard ( bh ) model and focus on the dynamics within the first band of the optical lattice .
in contrast to the vast literature which focuses on regimes around mott - insulator like phase transitions in the absence of an additional stark force , see e.g. @xcite , we concentrate on the bh model in the superfluid realm and in the presence of a static tilt .
we characterize the transition between the regular and the chaotic realm of the quantum spectra by a quantitative and systematical analysis based on statistical tests . in section [ sec : twobands ] we perturbatively include the decay to the second band via landau - zener like tunneling processes . the quantum spectrum of the latter , non - unitary problem is analyzed in section [ sec : results ] and found to essentially reproduce the properties of the purely one - band approximation down to rather small lattice depths .
the resulting decay rates for the interband tunneling strongly depend on the many - particle nature of the problem and are found to correlate with the transition to the quantum chaotic regime . as a consequence , signatures of many - body quantum chaos are predicted to be accessible to experiments with ultracold atoms over a broad range of parameters , in both `` horizontal '' transport along the lattice and in interband `` vertical '' transport .
our results are finally summarized in section [ sec : conclusion ] .
we briefly review the general hamiltonian for a system of spinless , interacting atoms in a quasi - one dimensional optical lattice subjected to a tunable tilting force @xmath0 .
we start out with the purely periodic problem , @xmath1 , in an optical potential of spacing @xmath2 , recoil momentum @xmath3 and typical kinetic energy @xmath4 . the optical potential and the kinetic energy form the single - particle bloch hamiltonian : @xmath5 the eigenfunctions are the bloch waves @xmath6 , labeled by the quasimomentum @xmath7 and the band index @xmath8 @xcite , with dispersion law @xmath9 . in the appendix [ sec : wannier ] we explain how to use the single - particle solutions @xmath10 to build a set of localized orbitals @xmath11 called wannier functions ( wf ) . in the limit of deep lattices
, the orbital @xmath11 goes to the wave function of the @xmath12th excited level for an harmonic potential centered on the @xmath13lattice site .
we use the wf to expand the field of the ultracold bosons : @xmath14 then we introduce a stark force @xmath0 that tilts the optical potential and a zero - range interaction between the atoms , parametrized by the scattering length @xmath15 . in a quasi - one dimensional optical lattice as realizable in experiments @xcite
the scattering length is derived from the true three - dimensional scattering length via a renormalization that accounts for the transverse confinement of the atomic wave functions @xcite and the physical dimension is then @xmath16=l^{-1}$ ] .
the hamiltonian in the second quantization is written as : @xmath17 \hat{\phi } ( x ) { \rm d}x \nonumber \\ & + & \frac{1}{2}\frac{4\pi a_{s } } { m } \int \hat{\phi } ^{\dag}(x ) \hat{\phi } ^{\dag}(x ) \hat{\phi } ( x ) \hat{\phi}(x ) { \rm d}x . \end{aligned}\ ] ] substituting the expansion eq . into eq .
we obtain the hamiltonian in terms of the creation and annihilation operators @xmath18 , @xmath19 , for a particle in the @xmath13th site and the @xmath8-th energy band of the lattice .
the number operator is @xmath20 .
we restrict the analysis to the first two bands , which can be addressed by experiments @xcite and that can be handled numerically without great difficulties .
the coefficients of the hamiltonian are given by integrals involving the wf : the exact computation of the wf outlined in appendix [ sec : wannier ] motivates the selection of the operators that are most relevant for @xmath21 .
we are left with the on - site energy @xmath22 , the kinetic energy @xmath23 , and the on - site interaction between atoms in the same band @xmath24 @xcite , for @xmath25 .
moreover we have on - site interaction between atoms in different bands @xmath26 and two transition operators @xmath27 , @xmath28 that are the subject of a detailed analysis in section [ sec : twobands ] .
the dimension @xmath29 of the hilbert space spanned by the fock states @xmath30 ( defined in appendix [ sec : floquetbloch ] ) , for a system of @xmath31 bosons distributed over @xmath32 lattice sites , occupying up to the @xmath33nd band of the periodic potential , is given by the combinatorial formula @xmath34 .
the typical number of lattice sites in experiments is @xmath35 and the filling - factor @xmath36 is of order unity @xcite , such that the exponential increase of @xmath29 with the system size limits any exact numerical approach to smaller systems , where we impose the cyclic boundary conditions @xmath37 .
the implementation of these conditions requires the system to be translationally invariant .
the stark potential @xmath38 , however , spoils the periodicity of the bloch hamiltonian eq . and produces localized ws eigenstates instead of traveling bloch waves @xcite .
we follow @xcite and proceed to eliminate the stark potential from the hamiltonian by changing to the interaction representation ( ir ) with respect to @xmath39 .
the hamiltonian in the ir , @xmath40 , is time - dependent , and the problem becomes conceptually more complicated .
we rescale the energies by @xmath41 , the lengths by @xmath2 , the momenta by @xmath42 and we make the substitutions @xmath43 , @xmath44 .
the on - site energies @xmath45 and the hopping amplitudes @xmath46 are given in eq . .
the interaction coefficients @xmath47 , @xmath48 , are proportional to the coupling constant @xmath49 and are given in eq . .
the `` dipole '' coefficient @xmath50 is given in eq . .
the hamiltonian of eq . ,
restricted to the first two bands of the lattice , finally reads in the ir : @xmath51 in the ir the hamiltonian is again symmetric for discrete translations in space and it has lost the time independence but it is periodic with the bloch period @xmath52 . the quasienergy levels @xmath53 of the fb operator for a system with @xmath54 , @xmath55 , in the subspace with @xmath56 .
the ws ladder is seen on the left panel , the perturbative splitting of the first rung is magnified in the lower panel . in the upper panel a single linear function of @xmath57 is added to @xmath58 to eliminate an overall winding trend , and allow a better visibility of the avoided crossings .
the parameters of the hamiltoniana are @xmath59 , @xmath60 , as used also in the subsequent figures .
] we assume that the initial state , for @xmath1 , is superfluid , characterized by the delocalization of the atoms over the entire lattice .
the critical conditions on the parameters , that enforce the superfluid phase in the present context @xcite is @xmath61 . following @xcite , in the present section we set the lattice depth @xmath62 , which gives @xmath63 and the interaction coefficient @xmath64 , with @xmath65 .
the object of the subsequent study is the evolution operator up to the bloch period @xmath66 , called floquet - bloch ( fb ) operator @xcite .
the results presented in the following confirm and extend the results of @xcite .
the discrete translational symmetry of the hamiltonian entails that the fb operator is a block - diagonal matrix in the basis of eq . , labeled by a many - body quasimomentum @xmath67 .
the dynamics of the atoms in the lattice is complex because many vectors take part in the time evolution of an arbitrary initial state .
the strong mixing of the basis vectors in time means that the evolution of a state is not bound to a small subspace of the total hilbert space ( contrary to @xcite ) , but , after a bloch period , the initial state spreads over the entire hilbert subspace with definite quasimomentum ( e.g. @xmath56 ) .
this is evidenced by the dependence on @xmath0 of the quasienergies @xmath68 , obtained from the eigenvalues @xmath69 of the fb operator . in the fig .
[ fig : oneband1 ] we show the quasienergy spectrum as the `` control parameter '' @xmath0 is varied . in the limit @xmath70
, we recover the standard bh model ( the analysis of the fb spectrum is here , however , not useful since the bloch period tends to infinity ) . in the regime
where the atomic interactions are negligible with respect to the lattice potential , the single particle bloch picture is adequate and the spectrum is simply a finite band .
for @xmath71 the single - particle ws ladder is found , i.e. a fan of energy levels @xmath72 , @xmath73 integer .
since the interactions are non - zero , the levels are split up and the first order perturbative effect on the ladder was computed in @xcite .
the central ws rung is split into levels which are proportional to the interaction energies of many atoms in a site , @xmath74 .
since the level splittings have the common factor @xmath75 , a collapse and revival of quasimomentum oscillations ( the fock space version of the single - particle bloch oscillations ) was predicted for this parameter range in @xcite . on the contrary ,
an irreversible decay of the quasimomentum oscillations was found in the range @xmath76 @xcite , where avoided crossings dominate the spectrum .
these are directly , experimentally observable consequences of the complex level structure presented in fig . [ fig : oneband1 ] . the distribution @xmath77 and the @xmath78 ( insets ) for the quasienergy spacings ( stairs ) , the wd ( solid ) and the poisson distribution ( dashed ) .
the parameters are @xmath79 , @xmath80 , @xmath81 ( left , regular ) and @xmath82 ( right , chaotic ) . ]
the presence of avoided crossings means that a strong mixing of the fock states is necessary to build the eigenstates of the system , and no set of quantum numbers can be assigned to individual levels as @xmath0 varies . in the region of parameters where the energy scales of the system are comparable in magnitude , we can characterize the spectrum in terms of the statistical distribution of the quasienergies spacings and by statistical measures for the eigenfunctions .
the latter have been intensively studied in @xcite . in the following
, we concentrate on the statistical behaviour of the quasienergies , which is closely linked to the behaviour of the open system studied in section [ sec : twobands ] . the probability @xmath83 that the magnitude of a given interval spacing @xmath84 is in @xmath85 $ ] is given by the poisson distribution @xmath86 @xcite for an uncorrelated spectrum in the regular regime .
strongly correlated quantum spectra , corresponding to the chaotic regime in our many - particle model , are well modeled by the wigner - dyson ( wd ) distribution for a circular orhogonal ensemble of random matrices @xcite : @xmath87 . in fig .
[ fig : oneband2 ] the probability distribution and the cumulative distribution function ( @xmath78 ) of the quasienergy spacings are shown for two paradigmatic values of @xmath0 .
the presence of avoided crossings in the chaotic case @xmath88 shows up as a depletion of small quasienergy spacings , and the probability to find a level crossing vanishes .
we improved the statistical description of the quasienergy spectrum with further analyses , shown in fig .
[ fig : oneband3 ] . in the panel (
a ) we quantify the convergence of the quasienergy spacings distribution to the wd profile , thus filling the gap between the two pictures of fig .
[ fig : oneband2 ] .
we computed the fb operator for several values of @xmath0 and confronted each spectrum with the wd distribution using a modified @xmath89 test , computed as follows .
each sequence of levels spacings was algorithmically binned to leave @xmath90 `` observed '' spacings in each bin @xmath91 @xcite . the `` expected '' values @xmath92 are the integrals of the theoretical distributions over the bins and the sum @xmath93 was calculated
. the values of @xmath94 are distributed according to a @xmath89 distribution with @xmath95 degrees of freedom and mean @xmath95 . the renormalized variable @xmath96\ ] ] is thus appropriate to compare several data sets , each binned optimally and independently , for @xmath97 , @xmath89 is in the range @xmath98 $ ] ( in the bulk of the original @xmath89 distributions before the transformation of eq . was performed ) , and the correspondent distribution of the spacings is well described by a wd profile . as the external force increases and the parameter @xmath99 diminishes , the larger values of @xmath89 ( lying in the tails of the original @xmath89 distributions ) indicate that the spectrum is not well characterized by a wd distribution .
the condition for the regular - chaotic transition can be directly read off from the quantitative statistical test results and corresponds , e.g. to @xmath100 .
statistical tests applied to the system of fig .
[ fig : oneband2 ] . ( a ) @xmath101like test of eq . .
( b ) @xmath102 test and ( c ) variance of number of levels ( with mean spacing normalized to 1 ) , for the chaotic ( circles ) and the regular ( pyramids ) case . in both panels
the solid and dot - dashed lines are the theoretical predictions for the wd and poisson statistics of energy levels , respectively . ]
we found that the profile of the quasienergy spacings distribution changes smoothly , but we used two statistical tests introduced in @xcite ( eq . ( 27 ) and ( 29 ) therein ) to show that the poisson and wd statistics are clearly identified at the borders of the transition .
[ fig : oneband3 ] ( b ) shows the explicit results for the @xmath102 function of @xcite whose predicted linear scaling @xmath103 for the poisson and @xmath104 for the wd expectation is confirmed . in fig .
[ fig : oneband3 ] ( c ) we show the variance @xmath105 @xcite of the number of levels @xmath106 found in a finite energy interval @xmath107 : @xmath108 ^ 2 \rangle_{e},\ ] ] where the average is taken over all the energies and we rescaled the spectrum so that the average number of levels per unit of energy @xmath109 is equal to unity . a linear and logarithmic scaling is predicted for a poisson and wd - like spectra , respectively .
the logarithmic behaviour clearly prevails over all energy ranges for the chaotic spectrum , only apart from oscillations which are indeed typical for samples of finite size ( see @xcite for details ) .
we develop a perturbative analysis of the two - bands system of eq . .
we consider the many - body dynamics within the ground band and the perturbative action of the operators that couple the fock states of the ground band to states in the second band . as a consequence of the interband terms
each fock state @xmath30 ( see appendix [ sec : floquetbloch ] ) suffers an energy shift @xmath110 , where @xmath111 is its decay width .
the hamiltonian is modified accordingly and becomes a non - hermitian effective hamiltonian for the ground band , that yields a non unitary fb operator . in the following ,
we compute the set of decay widths . in the next section
we then study the spectrum of this new fb operator .
we first define a basis of unperturbed states .
we choose to neglect the hopping in the lower band , where the wf are more strongly localized , so an unperturbed state projected on the hilbert space of @xmath31 particles in the ground band is a fock state @xmath112 . in the second band
we neglect the interactions , since in the perturbative approach only a few particles ( one or two in the following ) populate the excited levels .
so an unperturbed state projected on the second band is the solution of the one - particle ws problem @xcite , i.e. a localized wave function centred at site @xmath113 , written with the bessel function @xmath114 as : @xmath115 or @xmath116 , respectively . in the following ,
the occupation number @xmath117 for the @xmath13th site in the ground band is written @xmath118 .
the hamiltonian eq . contains two mechanisms that promote particles to the second band .
the first is a single - particle effect , a consequence of the external force , proportional to the dipole coupling @xmath50 between the wf of different bands .
the hamiltonian of the perturbation is @xmath119 the second perturbation is a many - body effect , describing two particles of the first band that collide and transform their interaction energy into kinetic energy , entering the second band together : @xmath120 the expectation value of @xmath121 , @xmath122 on the unperturbed states , equal to the first - order energy shift @xmath123 , is zero because the operators do not conserve the number of particles @xmath124 within the bands . let us focus on @xmath121 and compute its matrix element for the channel : @xmath125 the decay width at first - order is given by the fermi s golden rule and only the first term in eq .
gives nonzero contribution for the channel of eq . .
our result for the matrix element is : @xmath126 the kronecker @xmath127 functions act as a selection rule for the fock states that are coupled by the perturbation .
the tunneling mechanism does not include any income of energy from an external source , so the initial and final energies , @xmath128 must be equal as required by the golden rule .
the condition on the energy conservation is relaxed to account for the uncertainty @xmath129 of the unperturbed energy levels of the initial and final states . the energy uncertainty and
the level density function @xmath130 are derived from the perturbative action of the hopping operator of the first band that has been neglected so far .
we postpone the computation of these quantities to the end of the present section .
the relaxed energy conservation rule selects from eq .
the set @xmath131 of permitted decay channels @xmath132 parametrized by the two indices @xmath133 such that : @xmath134.\end{aligned}\ ] ] the last equation means that the energy @xmath135 required to promote a particle to the second band is supplied by the decrease of the repulsion energy ( proportional to @xmath136 ) and by the work of the force ( proportional to @xmath0 ) exerted on the promoted particle .
the total width @xmath137 for the decay via the allowed channels @xmath131 , is proportional to the square of the matrix element and to the level density @xmath130 given below in eq . .
we arrive at @xmath138 the perturbation @xmath122 is treated in a similar way , with the difference that two particles are promoted to the second band , and the position of the second stark state @xmath139 is an additional degree of freedom for the transition .
the decay channels are : @xmath140 the approximate energy matching equation selects a set @xmath131 of possible decay channels , parameterized by the three site indices @xmath141 : @xmath142.\end{aligned}\ ] ] we state the result for the decay width : @xmath143 with respect to eq . , the additional degree of freedom @xmath144 results in an extra summation extended over the ( infinite ) possible values of the difference @xmath145 .
this follows from the possibility to conserve the energy even if a particle is pushed very far , if the other particle is pushed almost equally far in the opposite direction . since the decay width eq .
depends on the product of two ( rapidly decaying ) bessel functions we apply the truncation @xmath146 , to reduce the formula to a finite form .
now we conclude the computation and derive the energy broadening @xmath129 of the fock states in the ground band necessary to implement the conditions of eq . and .
in the ground band , the unperturbed hamiltonian consists only of the on - site interaction operator : @xmath147 in the case of a single particle in a periodic potential the use of first order perturbation theory is wrong , as the second order of the perturbation theory diverges because of the exact degeneracy in energy of neighbouring sites , entailed by the translational symmetry of the lattice . on the contrary , in the present case
, the unperturbed hamiltonian is just a rough approximation of the true hamiltonian of the system and the remaining operators are supposed to remove the degeneracies , since the translational symmetry is broken by the external field in the ws picture .
the perturbation hamiltonian is given by @xmath148 and its matrix elements between fock states are @xmath149 ( a , b ) the probability distribution @xmath150 and ( c , d ) the @xmath78 for the _ logarithm _ of the decay widths @xmath151 in the regular regime [ ( a , c ) with @xmath152 and in the chaotic regime [ ( b , d ) with @xmath153 .
the size of the system is @xmath154 , @xmath155 .
the dashed line is the fit with a lognormal distribution .
the inset in panel ( d ) shows that the lognormal is not appropriate to fit the tails of the distributions in the chaotic regime . here and in the following figures , @xmath156 , @xmath157 , @xmath158 , @xmath159 , @xmath160 , @xmath161 , @xmath162 . ]
transitions are allowed between fock states that differ for one boson in two adjacent holes @xmath73 , @xmath163 .
the transition channels are written as @xmath164 , with @xmath165 and @xmath166 , and must fulfil the condition of the energy conservation : @xmath167 the total uncertainty is obtained by adding up the contributions from all the transitions and the summation over the allowed channels can be re - casted in a summation over the lattice sites . using the golden rule , each fock state is attributed the following energy uncertainty : @xmath168 finally , the level density @xmath130 around the unperturbed energy @xmath169 of a fock states @xmath170 is approximated by a rectangular profile , of width @xmath129 and area unity : @xmath171
( upper panel ) the @xmath101like test for the quasienergy spacings @xmath172 and ( lower panel ) the spread of the distribution of the decay widths @xmath151 .
the size of the system is @xmath154 , @xmath155 .
the dashed lines are a guide to the eye and suggest that the transition to the chaotic regime can be appreciated by looking at both , the real and the imaginary parts of the fb eigenvalues . ] the total width @xmath173 is now added to the single - band hamiltonian as a complex shift . given the translational symmetry of this hamiltonian , we added the shift to the diagonal in the representation of the cyclic basis @xmath174 in eq . , as @xmath175
the eigenvalues of the fb operator are no longer unitary and the quasienergies @xmath68 has a complex part @xmath176 .
we analyzed the decay rates @xmath177 _ along _ with the quasienergy spacings statistics to study how the dynamics within the first band influences the coupling to the second band .
we first reported the distribution @xmath178 in @xcite . here
we refine our analysis and we first focus on @xmath179 , with @xmath180 , shown in fig . [
fig : results1 ] for some paradigmatic cases .
the widths are much smaller than unity , consistently with a perturbative approach of the system , yet the lattice potential is only @xmath181 to increase the spread of the stark state in the second band .
moreover , the decay channels are activated by an increase of the interaction energy , which can be experimentally achieved by acting on the transversal confining potential @xcite of the quasi - one dimensional lattice , or by a feshbach resonance @xcite . in this section ,
@xmath182 used in @xcite is multiplied by a factor of order @xmath183 , a value that is still well within the experimental possibilities @xcite .
( color online ) ( a - c ) l.h.s of eq . in the regular ( a ) and in the chaotic regime ( b , c ) , for a system with @xmath154 , @xmath155 .
the dashed area in panels ( b , c ) shows the part of the histogram where the total amount of points is about @xmath184 , less then @xmath183% of the full sample . the red solid line is the linear fit to the profile in the scaling region .
( d ) the exponent of the power law @xmath185 as obtained from the linear fits shown in ( a - c ) , for @xmath154 , @xmath155 ( black circles ) , @xmath154 , @xmath55 ( red squares ) , @xmath186 , @xmath80 ( blue diamonds ) , @xmath186 , @xmath55 ( green pyramids ) . to show together different data sets we horizontally translated the first by @xmath187 and the last by @xmath188 . ] in fig .
[ fig : results1 ] we compare the distribution of the decay widths for two values of @xmath0 that belong to the regular ( a , c ) and the chaotic region ( b , d ) .
the difference in the average decay width @xmath189 is due to the improved energy matching provided by a stronger external field @xmath0 , that supplies the necessary energy to promote particles to the second band . for the parameters of fig .
[ fig : results1 ] , the single - particle landau - zener formula @xcite gives @xmath190 \sim 10^{-20}$ ] , @xmath191 for ( a , b ) .
then we see that there are regimes where the many - body interactions affect substantially the single - particle tunneling rates and _ can not _ be neglected .
moreover , even mean - field treatments of the landau - zener tunneling @xcite can not account for several decay channels .
the _ bulks _ of the distributions , both in the regular and in the chaotic regime , are appropriately fit by a lognormal profile @xmath192^{2}/2\delta\lambda^{2}\}/(\lambda-\lambda_{\rm min})$ ] .
we notice that the spread @xmath193 is reduced in the chaotic case , where the fock states are strongly mixed by the dynamics and a fast decaying fock state ( in the @xmath174 representation ) could be a privileged decay channel for _ many _ eigenstates of the system .
many eigenstates then shared similar decay widths and their statistical distribution would be thinner .
following this reasoning , we interpret the thinner distributions found in the panels ( b , d ) of fig .
[ fig : results1 ] as a signature of the strongly correlated dynamics .
this picture is supported by fig .
[ fig : results2 ] , where the dependence on @xmath0 of the spread @xmath193 ( lower panel ) is confronted with the regular - chaotic transition evidenced by the @xmath101like test of eq .
( upper panel ) .
the shrinking of the decay widths distribution goes along with the transition , though the precise determination of a transition point is precluded by a substantial amount of noise .
the finite size of the samples that can be managed numerically accounts for the noise , as we verified that the profile becomes sharper while increasing the size of the system .
moreover , since the average decay width decreases with smaller @xmath0 , we need a more precise and hence a more time - expensive numerical computation to determine @xmath193 as we enter the chaotic regime . finally , we found in @xcite that the tails of the distributions in the chaotic case follow the expected power - law for the diffusive regime of an open quantum chaotic system . in our case ,
the opening of the ground band subsystem is defined by the interband coupling , which in a sense attaches `` leads '' to _ all _ lattice sites _ within _ the sample .
indeed , the inset of fig .
[ fig : results1 ] ( d ) shows that the lognormal is not a good fit to the tails in the chaotic regime : the distributions transform to a power - law @xmath185 in close analogy to the transition from anderson - localized to diffusive dynamics in open disordered systems @xcite .
in particular we found @xmath194 , in accordance with the general results of @xcite . in the fig .
[ fig : results3 ] , panel ( d ) we show that this value is indeed peculiar to the chaotic regime , and strong fluctuations with @xmath0 mean that the exponent @xmath8 is not defined within the transition region . here
we evaluate the integrated profile @xmath195 and fit the latter to reduce the statistical fluctuations due to the finiteness of our samples .
the finite - size effects are more relevant in the chaotic case ( panels b , c ) , because the reduced spread of the distribution makes the fit sensible to the few points that fall in the further part of the tail ( shaded in the pictures ) .
the largest analyzed sample has only @xmath196 energy levels , yet the uncertainty on @xmath197 on the chaotic side of the transition is less than @xmath183% .
we studied the problem of a many - body atomic system in a tilted optical lattice , using a statistical analysis of the complete quantum spectrum .
we extended previous work @xcite and verified thoroughly the transition from a regular to a quantum chaotic spectrum , found in @xcite as the stark field is varied .
the transition takes place when the tunneling amplitude in the lattice becomes comparable to the interaction energy of the atoms . in this regime
it is not possible to find a set of quantum numbers that decompose the spectrum into subspaces not mixed by a change of the stark field .
because of the strong mixing of the energy eigenstates , many avoided crossings are found in the spectrum and the statistical analysis of the energy spacings show a characteristic depletion of small spacings - a signature of quantum chaos .
we derived a reasonable two - bands model , on the basis of which we opened the ground band subsystem by a perturbative coupling of the fock states to excited stark states .
we obtained the set of decay rates from the complex - valued quasienergies of the fb operator .
we analyzed the real part of the quasienergies and verified that the transition from the regular to the chaotic regime is still visible and not much modified with respect to the one - band system .
moreover we analyzed the statistical distribution of the imaginary part of the quasienergies , i.e. the decay rates of the states in the ground band .
we found that the distributions of the decay rates become thinner as the regular - chaotic transition is crossed .
we reckon that thinner distributions of the decay rates are a signature of the complex dynamics , where a few strongly decaying states act as leading decay channels .
the statistical characterization of the tunneling rates could be used to compute the expected atomic current from the ground band to the excited band of the lattice , thus providing an experimental probe for the regular - chaotic transition .
time - dependent observables could possibly be computed also with advanced mean - field techniques as reported in ref .
@xcite , as long as no full spectral information of the many - body system is wanted .
of course , a deeper investigation is desirable to understand on quantitative grounds the full interplay of the dynamics within the first band and the decay towards higher bands , and a more detailed analysis of the interband coupling will be worthwhile in a full - blown model in which at least two bands are completely included .
our results are a first step in this direction of studies of regimes in which both `` horizontal '' and `` vertical '' quantum transport are simultaneously present and influence each other in a complex manner .
we thank alexey ponomarev , gregor veble , and andrey kolovsky for lively discussions on the wannier - stark system and quantum chaos and yan fyodorov for bringing ref .
@xcite to our attention .
we are grateful to the centro di calcolo , dipartimento di fisica , universit di pisa , for providing cpu and kind assistance , and for support by miur - prin and eu - olaqui .
s. w. furthermore acknowledges support within the framework of the excellence initiative by the german research foundation ( dfg ) through the heidelberg graduate school of fundamental physics ( grant number gsc 129/1 ) .
the coefficients of the hamiltonian eq . are given by integrals of the wf @xmath198 , once the expansion of eq .
is substituted into eq . .
the wf are defined as : @xmath199 from the definition it follows that @xmath200 , so that all the bloch waves for a band can be computed starting from a _ single _ wannier function @xmath201 .
we computed the wf following the method introduced in @xcite .
a bloch wave is factorized as @xmath202 , where the periodic function @xmath203 is expanded in a truncated basis of momenta using multiples of @xmath204 only : @xmath205 using a gauge transform ( @xmath206 ) , an effective hamiltonian for the periodic @xmath207 is derived from eq . .
we solved the effective schrdinger equation as a linear system for @xmath208 with dispersion law @xmath9 as eigenvalues .
we obtained the on - site energies and the hopping amplitudes from the fourier transform of @xmath9 : @xmath209 the wf is finally computed from the inversion of eq . .
the relative phase of the eigenvectors @xmath208 , for different quasimomenta , is _ not _ unique . for
the simple case of a sinusoidal lattice potential , the correct choice of the phases can be inferred from @xcite : @xmath210 this phase choice guarantees the correct inversion symmetry @xmath211 of the wf of the first and of the second band , respectively .
the interaction coefficients for eq . read : @xmath212 the wf are an orthogonal set of functions , so that different bands are decoupled in the one - body dynamics .
the tilting potential @xmath38 has nonvanishing `` dipole '' matrix elements that couples adjacent bands : @xmath213
the fock states @xmath30 are defined by the sequence of occupation numbers @xmath214 for the @xmath32 sites of the lattice , in each band @xmath8 .
the cyclic boundary conditions , @xmath215 allow us to define a shift operator @xmath216 @xcite that translates the occupation numbers of a fock state @xmath217 : @xmath218 the operator @xmath216 naturally decomposes the fock space into equivalence classes of vectors generated by repeated application of @xmath216 onto a `` seed '' vector @xmath219 with @xmath220 such that @xmath221 for all the vectors @xmath222 in the class . a new basis @xmath223 can be introduced for which the many - body quasimomentum @xmath224 ( @xmath225 ) , supplies a convenient label : @xmath226 with the obvious advantage that we can diagonalize separately each block of dimension @xmath227 .
this decomposition not only leads to a substantial numerical simplification but is also of dynamical relevance @xcite .
moreover , we exploited that the hamiltonian matrix is _ sparse _ and we verified that the fraction of the nonzero entries is @xmath228 in the limit of large hilbert spaces @xmath229 .
the column @xmath230 of the fb operator coincides with the column of the coefficients of the basis state of index @xmath230 , evolved in time up to one bloch period .
we used a fourth - order runge - kutta time integrator with adaptive stepsize , tuned in precision by the upper bound @xmath231 of the estimated one - step error @xcite .
the value of @xmath231 was chosen to suppress , up to @xmath232 , the well - known exponential instability of the runge - kutta method applied to the schrdinger equation .
the quantity @xmath233 was used to compare different spectra @xmath234 , @xmath235 and we verified the consistency of our computations , finding a power - law self - convergence @xmath236 for a sequence of tests with increasing required precision , @xmath237 . the achieved precision scales with cpu time @xmath238 as @xmath239 .
a precision up to @xmath240 was necessary to reliably compute the small complex part in the eigenvalues of the fb operator , analyzed in section [ sec : results ] . | we study a model for ultracold , spinless atoms in quasi - one dimensional optical lattices and subjected to a tunable tilting force .
statistical tests are employed to _
quantitatively _ characterize the spectrum of the floquet - bloch operator of the system along the transition from the regular to the quantum chaotic regime .
moreover , we perturbatively include the coupling of the one - band model to the second energy band .
this allows us to study the landau - zener interband tunneling within a truly many - body description of ultracold atoms .
the distributions of the computed tunneling rates provide an independent and experimentally accessible signature of the regular - chaotic transition . |
`` information '' is by its nature _ broadcastable_. what happens when information is quantum and we need to distribute it among many users ? indeed , this may be useful in all situations where quantum information is required in sharable form , e. g. in distributed quantum computation @xcite , for quantum shared secrecy @xcite , and , generally , in quantum game - theoretical contexts @xcite .
however , contrarily to the case of classical information , which can be distributed at will , broadcasting quantum information can be done only in a limited fashion .
indeed , for pure states ideal broadcasting is equivalent to the so - called `` quantum cloning '' , which is impossible due to the well - known `` no - cloning '' theorem @xcite ( see also @xcite ) .
the situation is more involved when the input states are mixed , since broadcasting can be achieved with an output joint state which is indistinguishable from the tensor product of local mixed states from the point of view of individual receivers .
therefore , the no cloning theorem can not logically exclude the possibility of ideal broadcasting for sufficiently mixed states . in ref .
@xcite it was proved that perfect broadcasting is impossible from a single input copy to two output copies for an input set of non mutually commuting density operators .
this result was then considered ( see refs . @xcite and @xcite ) as an evidence of the general impossibility of broadcasting mixed states drawn from a non - commuting set in a more general scenario , where @xmath1 equally prepared input copies
are broadcast to @xmath2 users . however , for sufficiently many input copies @xmath3 and sufficiently mixed input states the no - broadcasting theorem does not generally hold @xcite , and for input mixed states drawn from a noncommuting set it is possible to generate @xmath2 output local mixed states which are identical to the input ones , by a joint correlated state
actually , as proved in ref .
@xcite , it is even possible to partially purify the local state in the broadcasting process , for sufficiently mixed input states .
such a process of simultaneous purification and broadcasting was named _
superbroadcasting_. for qubits , the fully covariant superbroadcasting channel that maximizes the output purity ( i. e. the length of the output bloch vectors of local states ) when applied to input pure states coincides with the optimal cloning map @xcite
. the possibility of superbroadcasting does not increase the available information about the original input state , due to unavoidable detrimental correlations among the local broadcast copies , which do not allow to exploit their statistics ( a similar phenomenon was already noticed in ref .
@xcite ) . essentially , superbroadcasting transfers noise from local states to correlations . from the point of view of single users
, however , the protocol is a purification in all respects , and this opens new interesting perspectives in the ability of distributing quantum information in a noisy environment , and deserves to be analyzed in depth . for qubits , it has been shown that for universal superbroadcasting is possible with at least @xmath4 input copies @xcite .
is this the absolute minimum number for superbroadcasting , or does it hold only for this particular set of input states ? in this paper
we will show that , indeed , for equatorial mixed states of qubits the minimum number is @xmath5 .
however , for smaller non - commuting sets of qubit states the possibility of superbroadcasting with only @xmath6 input copies is still an open problem ( for larger dimension @xmath7 it is possible to superbroadcast also for @xmath6 , see e.g. ref .
@xcite ) . we want to point put that clearly there are limitations to superbroadcasting
. the input state must be indeed sufficiently mixed , since pure states can not be broadcast by the no cloning theorem .
however , states with a pretty high purity can sill be superbroadcast , e. g. for universally covariant superbroadcasting @xcite from @xmath4 to @xmath8 it is possible to superbroadcast states with a bloch vector length up to 0.787 ( 0.935 for the phase covariant case ) .
one can achieve superbroadcasting with even higher input purity for increasing @xmath3 , approaching unit input bloch vector length in the limit of infinitely many input copies .
there are also some limitations in the absolute number of output copies for which one can achieve superbroadcasting .
e. g. in the universal case , for @xmath4 input copies one can superbroadcast up to @xmath9 output copies , for @xmath10 up to @xmath11 , and for @xmath12 up to infinitely many .
the output purity is clearly decreasing versus @xmath13 . in this paper
we will further analyze all limitations to superbroadcasting also for the case of equatorial input qubits .
regarding the possibility of achieving superbroadcasting experimentally , the first route to explore is to use the same techniques as for purification @xcite , since the superbroadcasting map for @xmath2 generalizes the purification map using the same protocol @xcite .
this transformation involves a measurement of the total angular momentum of the qubits , then an optimal werner cloning @xcite in the universal case ( or an optimal phase covariant cloning @xcite in the phase covariant case ) .
another possibility is to use the methods of ref .
@xcite in order to classify all possible unitary realizations , and then seek for experimentally achievable ones using current technology . in this paper
we present phase covariant superbroadcasting , and also give a complete derivation of the universal superbroadcasting map , presented in ref .
the two maps are derived in a unified theoretical framework .
the paper is organized as follows . in section [ s : symm - broad ]
we introduce some preliminary notions regarding symmetric covariant maps . in sects . [
s : univ ] and [ s : pc ] we give a complete derivation of the optimal broadcasting maps in the universal case and in the phase covariant case respectively . in sect .
[ sec : concurrence ] we study the entanglement properties of the states of two copies at the output of the universal and the phase covariant broadcasting maps . finally , in sect .
[ s : conc ] we summarize and comment the main results of this paper . at the end of the paper
we report the details of the calculations needed to derive the results presented in three appendices .
in this section we introduce in a unified theoretical framework some preliminary concepts that will be employed to describe covariant symmetric qubit broadcasting maps .
these concepts will be then specified to the universal and the phase covariant cases in the subsequent sections .
a main tool we will extensively use in deriving the optimal maps is the formalism of the choi - jamiokowski isomorphism @xcite between completely positive ( cp ) maps @xmath14 from states on the hilbert space @xmath15 to states on the hilbert space @xmath16 , and positive bipartite operators @xmath17 on @xmath18 .
such an isomorphism can be specified as follows @xmath19,\ ] ] where @xmath20 is the non normalized maximally entangled state @xmath21 in @xmath22 , @xmath23 gives the identity transformation and @xmath24 denotes the transposition of the operator @xmath25 on the same basis @xmath26 used in the definition of @xmath20 . in terms of @xmath17 ,
the trace - preservation condition for the map @xmath14 reads @xmath27=\openone_\sh\,.\ ] ] suppose that the map @xmath14 is covariant under the action of a group @xmath28 . in this case
the covariance property is reflected to the form of the operator @xmath17 by the following correspondence @xmath29=0.\ ] ] in the above expression @xmath30 and @xmath31 are the unitary representations of @xmath32 on the input and output spaces respectively , while @xmath33 denotes complex conjugation on the fixed basis @xmath26 . in this framework
it is also possible to study group - invariance properties of the map @xmath14 in terms of the operator @xmath17 . in this case
we have the following equivalences @xmath34=0,\ ] ] and @xmath35=0.\ ] ] the above expressions refer to invariance properties on the input and output spaces respectively . in the following
we will consider maps @xmath36 from states of @xmath3 qubits to states of @xmath13 qubits , namely cp maps from states on @xmath37 to states on @xmath38 .
we will consider in particular symmetric broadcasting maps , namely transformations where all receivers get the same reduced state .
the figures of merit which are commonly used are invariant under permutations of the output copies , and this allows to assume that the output state of a broadcasting map is permutation invariant without loss of generality .
moreover , since the input consists in @xmath3 copies of the same state , there is no loss of generality in requiring that the map is also invariant under permutations of the input copies . these two properties , according to eqs .
( [ invariance1 ] , [ invariance2 ] ) , can be recast as follows @xmath39=0,\qquad\forall\sigma,\tau , \label{perminv}\ ] ] where @xmath40 and @xmath41 are representations of the output and input copies permutations , respectively .
notice that permutations representations are all real , hence @xmath42 .
a useful tool to deal with unitary group representations @xmath30 of a group @xmath28 on a hilbert space @xmath15 is the wedderburn decomposition of @xmath15 @xmath43 where the index @xmath44 labels the equivalence classes of irreducible representations which appear in the decomposition of @xmath30 . the spaces @xmath45 support the irreducible representations and @xmath46 are the multiplicity spaces , with dimension @xmath47 equal to the degeneracy of the @xmath44-th irreducible representation .
correspondingly the representation @xmath30 decomposes as @xmath48 where @xmath49 is shorthand for @xmath50 . by schur
s lemma , every operator @xmath25 commuting with the representation @xmath30 in turn decomposes as @xmath51 in the case of permutation invariance , the so - called schur - weyl duality @xcite holds , namely the spaces @xmath46 for permutations of @xmath13 qubits coincide with the spaces @xmath45 for the representation @xmath52 of @xmath53 , where @xmath30 is the defining representation .
in other words , a permutation invariant operator @xmath54 can act non trivially only on the spaces @xmath45 , namely it can be decomposed as @xmath55 the clebsch - gordan series for the defining representation of @xmath53 is well - known in the literature ( see for example @xcite ) , and its wedderburn decomposition is given by @xmath56 where @xmath57 , @xmath58 equals 0 for @xmath13 even , 1/2 for @xmath13 odd , and @xmath59 in the case of the broadcasting maps the hilbert space @xmath18 on which the operator @xmath17 acts , supports the two permutation representations corresponding to the output and input qubits permutations .
therefore it can be decomposed as @xmath60 by rearranging the factors in the above expression , we can recast the decomposition in a more suitable way , namely @xmath61 the operator @xmath17 , in order to satisfy the permutation invariance property , according to eq .
( [ eq : schur - weyl - dual - form ] ) , can be written in the following form @xmath62 where the operators @xmath63 act on @xmath64 . moreover , in order to fulfill the requirements of trace preservation and complete positivity , the operators @xmath63 must satisfy the constraints @xmath65=\frac{\openone_{2l+1}}{d_j } , \label{cpt}\ ] ] where @xmath66 denotes the partial trace performed over the space @xmath67 in the @xmath68-th term of the decomposition in eq . , and @xmath69 is shorthand for @xmath70 .
we have now all the tools to study symmetric qubits broadcasting devices . in this work
we are interested in the case of covariant broadcasting maps , which in general have to fulfill also the following covariance condition under the representations @xmath71 and @xmath72 of a group @xmath28 ( see eq .
( [ eq3 ] ) ) @xmath73=0.\ ] ] the above condition gives a further constraint on the form of the operators @xmath63 in eq .
( [ symmbro ] ) . actually , the group @xmath31 is in general just a subgroup of the defining representation @xmath30 of @xmath53 , and therefore the representation @xmath74 acts non trivially only on the subspaces @xmath64 , which are the ones supporting the operators @xmath63 . in the next sections we will consider two interesting cases , namely @xmath75 and @xmath76 , corresponding to universal and phase covariant broadcasting respectively , and
we will see how the form of the operators @xmath63 depends on the particular choice of the considered covariance group .
in addition to the wedderburn decomposition and the related schur - weyl duality reviewed above , another useful tool we will extensively use in the following is a convenient decomposition of an @xmath3-partite state of the form @xmath77 , representing @xmath3 qubits all prepared in the same generic state @xmath78 . in appendix [ appendixaa ]
we report the complete derivation of the following identity , which was originally presented in @xcite , @xmath79 for the sake of simplicity , in the above expression we considered density operators of the form @xmath80 , @xmath81 , namely qubit states whose bloch vector ( of length @xmath82 ) is aligned along the @xmath83 axis , and consequently the states @xmath84 are eigenstates of the operator @xmath85 in the @xmath68 representation , namely @xmath86 .
notice actually that the _ total _ angular momentum component @xmath85 of @xmath3 qubits is clearly permutation invariant and therefore it can be written as @xmath87 we want to point out that the decomposition ( [ eq : rho - decomposition ] ) holds for any direction of the bloch vector , provided that the eigenstates of @xmath85 in eq .
( [ eq : rho - decomposition ] ) are replaced by the eigenvectors of the angular momentum component along the direction of the bloch vector in the single qubit state @xmath78 .
we will prove in appendix c that the single - site output copy @xmath88 of a covariant broadcasting map commutes with the input density operator @xmath78 . in order to quantify the performance of the broadcasting map @xmath36 and to judge the quality of the single - site output density operator @xmath89 $ ] we will then evaluate the length @xmath90 of its bloch vector , namely @xmath91=r^\prime\ , .
\label{singsitra}\ ] ] notice , moreover , that the length of a bloch vector @xmath82 is simply related to the purity @xmath92 $ ] of the density operator @xmath78 as @xmath92=(1+r^2)/2 $ ] .
therefore , maximizing the output bloch vector length @xmath93 is equivalent to maximizing the output single - site purity . notice also that so far we can not exclude that the input and output bloch vectors @xmath82 and @xmath93 are antiparallel , and this just implies that @xmath90 can range from @xmath94 to @xmath95 .
we will now show how to evaluate @xmath93 according to eq .
( [ singsitra ] ) , which is the main quantity that we will consider in the next sections in the particular cases of universal and phase covariant broadcasting .
the trace in eq . can be computed by considering that the global output state @xmath96 of the @xmath13 copies is by construction invariant under permutations , hence @xmath97=\tr\left[\left(\sigma_z\otimes\openone^{\otimes m-1}\right)\sigma\right]=\frac1{m!}\tr\left[\sum_\sigma\pi_\sigma\left(\sigma_z\otimes\openone^{\otimes m-1}\right)\pi_\sigma \sigma\right]\,.\ ] ] the last term on the r.h.s . of eq . ( [ eq : rprime - valutaz ] ) contains a sum over the @xmath98 possible permutations of the @xmath13 output qubits .
notice that @xmath99 where the operator @xmath100 acts as @xmath101 on the @xmath102-th qubit and identically on the remaining qubits .
now , by exploiting the permutation invariance of @xmath103 , we can write @xmath104 and clearly @xmath105\,.\ ] ] the explicit expression of @xmath93 in the universal case will be derived in sect .
[ s : univ ] . in the phase covariant case
, we will see that it is more convenient to take @xmath78 diagonal on the @xmath106 eigenstates .
the above formula in this case is just substituted by @xmath107\,\ ] ] and will be explicitly calculated in sect .
[ s : pc ] .
we want to stress that maximization of the figure of merit @xmath93 allows to optimize the fidelity criterion as well .
in fact , for @xmath108 the two criteria coincide , whereas for @xmath109 one can always achieve unit fidelity by suitably mixing the output state with optimal @xmath93 and the maximally mixed one . on the other hand , direct maximization of fidelity
is not analytically feasible , since fidelity is a concave function over the convex set of covariant maps , whence it is not maximized by extremal maps .
finally , we want to mention that in the next sections we explicitly maximize the scaling factor for @xmath3 inputs and @xmath13 outputs @xmath110 , which can be referred to as _ shrinking factor _ or _ stretching factor _ , depending whether it is smaller or greater than 1 , respectively .
it is obvious that this maximization is equivalent to maximizing @xmath93 .
superbroadcasting corresponds to the cases where @xmath111 .
in this section we will give the explicit derivation of the optimal universal broadcasting maps .
starting from the general broadcasting map described in the previous section we have to impose in this case the additional constraint @xmath112=0\,,\ ] ] where @xmath30 is the defining representation of the group @xmath53 . for the defining representation
@xmath30 the following property holds @xmath113 by exploiting such a property , the commutation relation ( [ univ1 ] ) can be written more conveniently as follows @xmath114=0\ , , \label{universales}\ ] ] where @xmath115 .
the complete positivity and trace - preservation constraints in terms of the operator @xmath116 are then equivalent to @xmath117=\openone_\sh\,.\ ] ] upon defining @xmath118 , the constraints for complete positivity and trace preservation are now given by the following conditions on the operators @xmath119 @xmath120=\frac{\openone_{2l+1}}{d_j}.\ ] ] by exploiting the fact that the clebsch - gordan series for @xmath64 is just @xmath121 , we can write @xmath122 notice that this is not the wedderburn decomposition , since not all the subspaces @xmath123 support inequivalent representations . however , the wedderburn decomposition can be recovered by a suitable rearrangement that takes into account the repetitions of the same representation @xmath124 . using the decomposition we can formulate the constraint in terms of the operators @xmath119 as follows @xmath125 where , by complete positivity , the coefficients @xmath126 are real and positive , and @xmath127 is the projection of the space @xmath64 onto the @xmath124 representation , satisfying @xmath128=\frac{2j+1}{2l+1}\openone_{2l+1},\qquad\tr_l\left[p^j_{j , l}\right]=\frac{2j+1}{2j+1}\openone_{2j+1}.\ ] ] the set of projectors @xmath127 is clearly orthogonal .
the trace - preservation constraint ( [ trace - pres - abovo ] ) can now be written as @xmath129 which is equivalent to the conditions @xmath130 along with the complete positivity constraint @xmath131 , eq .
( [ eq : polytope ] ) defines a convex polyhedron whose extremal points are classified by functions @xmath132 and @xmath133 @xmath134 the classification of symmetric universally covariant maps is then completely determined in terms of the vectors @xmath135 and @xmath136 , whose elements can range from @xmath58 to @xmath137 and from @xmath138 to @xmath139 , respectively .
extremal maps then correspond to the following form for the operators @xmath116 @xmath140 the optimization of the figure of merit @xmath90 or , equivalently , of the scaling factor @xmath141 can be obtained by explicit calculation from eq .
( [ eq : rprime ] ) . the output state @xmath103 of the broadcasting map applied to an input state @xmath77 can be represented as @xmath142=\tr_\sh[\openone^{\otimes m}\otimes{\tilde\rho}^{\otimes n}\ s]\ , , \label{maponinp}\ ] ] where @xmath143 denotes the orthogonal complement of @xmath78 , which just corresponds to the change @xmath144 ( or , equivalently , @xmath145 ) . using the decomposition in eq .
( [ eq : rho - decomposition ] ) for @xmath146 and the form ( [ s ] ) for the operator @xmath116 , we can express eq . as follows @xmath147\otimes\openone_{d_{j_l}}.\ ] ] we can now use eq .
( [ eq : rprime ] ) to evaluate the scaling factor , namely @xmath148.\ ] ] in appendix [ appendixb ] we report the explicit calculation of @xmath149 and we show that it can be written in the following form @xmath150 where @xmath151 since @xmath152 , the sum @xmath153 in eq .
( [ eq : univ - pr - to - opt ] ) is always negative .
therefore , the function @xmath149 is maximized by the choice of @xmath136 and @xmath135 minimizing @xmath154 , which clearly implies @xmath155 .
the form of the coefficient @xmath156 for @xmath157 is given by @xmath158 whereas for @xmath159 we have @xmath160 in both cases @xmath154 is minimized by choosing the maximum value of @xmath135 , and therefore the maximum scaling factor is achieved by @xmath161 .
for @xmath2 the optimal value of the figure of merit is then univocally determined by the value of the function @xmath154 @xmath162 while the optimal scaling factor is given by @xmath163 the corresponding output state takes the form @xmath164 where @xmath165 denote the clebsch - gordan coefficients .
as mentioned in the previous section , in appendix [ appendixa ] we prove that the single - site reduced output state @xmath166 $ ] commutes with @xmath101 , hence @xmath149 is definitely a scaling factor .
two interesting cases we will consider in the following are the ones with @xmath167 and @xmath168 , for which the scaling factor takes the explicit forms @xmath169 the function @xmath170 is plotted in fig .
[ prplotz ] for @xmath3 ranging from 10 to 100 in steps of 10 .
we can see that for a suitable range of values of @xmath82 the scaling factor is larger than one .
this corresponds to a broadcasting process with an increased single - site purity at the output with respect to the input .
this phenomenon occurs for @xmath171 .
in this case @xmath141 is actually a stretching factor , and we call such a phenomenon superbroadcasting . the maximum value of @xmath82 such that it is possible to achieve superbroadcasting will be referred to as @xmath172 and it is a solution of the equation @xmath173 versus @xmath82 for the universal broadcasting , for @xmath167 and @xmath3 ranging from 10 to 100 in steps of 10 . notice that there is a wide range of values of @xmath82 such that @xmath174 , corresponding to superbroadcasting.,width=340 ] it is clear that the optimal scaling factor for fixed @xmath3 is a non increasing function of @xmath13 .
actually , by contradiction , suppose that the map with @xmath175 output copies has a higher purity than the optimal map with @xmath13 copies
. then one could trace over @xmath176 copies from the former map , and he would obtain a map with @xmath13 output copies with purity higher than the optimal , which is obviously absurd .
this implies that in general @xmath177 , and for large values of @xmath176 superbroadcasting may not be possible anymore .
the maximum @xmath13 such that superbroadcasting can be achieved for @xmath3 input copies will be referred to as @xmath178 .
it turns out that , apart from the values @xmath179 for which we have @xmath180 and @xmath181 , for @xmath182 one has @xmath183 , namely superbroadcasting is possible for any number of output copies . in fig .
[ rstar ] we report the values of @xmath184 and @xmath185 for @xmath186 . by a numerical analysis
we have evaluated the power laws for the two curves , which turn out to be in good agreement with @xmath187 and @xmath188 , respectively .
we want to point out that for input pure states ( @xmath189 ) only the term with @xmath190 in the expression ( [ s ] ) is significant .
the optimal map then corresponds to the optimal universal cloning for pure states derived in @xcite . and
@xmath185 for @xmath186 .
the upper line , corresponding to @xmath191 , has a power law @xmath188 .
the lower line , corresponding to @xmath167 , has a power law @xmath187.,width=340 ]
in this section we study the case of symmetric phase covariant broadcasting , where we restrict our attention to input states lying on an equator of the bloch sphere , say the @xmath192-plane .
the equatorial qubit density operator in this case has the explicit form @xmath193 .
the starting point , as in the case of universal broadcasting , is the requirement of permutation invariance ( [ perminv ] ) for input and output copies , that leads to the form ( [ symmbro ] ) of the operator @xmath17 .
moreover , in this case we demand covariance under the action of the group of rotations along the @xmath83-axis @xmath194 .
we proceed analogously to the case of universal broadcasting . by imposing the covariance condition for the map we require invariance of the @xmath17 operator , namely @xmath195=0
$ ] . by exploiting the wedderburn decomposition ( [ eq : wedder - for - u ] ) for the operator @xmath196 ,
namely @xmath197 the phase covariance requirement corresponds to the following additional condition for the operators @xmath63 @xmath198=0,\qquad\forall j , l,\ ] ] where @xmath199 is defined according to eq .
( [ jz ] ) . a convenient form for the operators @xmath63 satisfying eq .
( [ eq : rjl - covariance ] ) is the following @xmath200 when @xmath201 , and @xmath202 when @xmath203 .
notice that there are two more running indices with respect to the universal case ( [ onlyonerunning ] ) . the index @xmath204 in eq .
( [ eq : rjl1 ] ) simply allows for off - diagonal contributions in the operator @xmath63 , while we will see that the index @xmath205 , which labels equivalence classes , is related to the direction of the reduced output state bloch vector . in particular we will show that , in order to get an equatorial output , the operators @xmath63 have to be symmetric in @xmath205 , in the sense that @xmath206 .
notice also that @xmath205 takes integer values when @xmath207 is even and half integer values when @xmath207 is odd .
the trace - preservation condition ( [ cpt ] ) now reads @xmath208 and , analogously to the universal case , the fact that the operators @xmath63 are diagonal with respect to the indices @xmath68 s and @xmath205 s implies that the extremal points are classified by functions @xmath209 and satisfy @xmath210 we will now compute the output density operator and the scaling factor for @xmath0 phase covariant broadcasting maps . without loss of generality ,
let us now consider an input state @xmath78 oriented along the @xmath211-axis , namely @xmath212 .
the density operator @xmath77 can then be decomposed , analogously to eq .
( [ eq : rho - decomposition ] ) , as @xmath213 where @xmath214 is the eigenvector of @xmath215 corresponding to the eigenvalue @xmath216 . in the following ,
eigenvectors without explicit specification of the superscript axis , such as @xmath84 , are intended to be along the @xmath83-axis , namely @xmath217 . according to eq .
( [ eq : choi - jam ] ) , the density operator @xmath103 on @xmath218 , describing the output state of the @xmath13 copies , can be written as @xmath219=\tr_\sh\left[\left(\openone^{\otimes m}\otimes\rho^{\otimes n}\right)r\right]\\ & = ( r_+r_-)^{n/2}\bigoplus_{j = j_0}^{m/2}\sum_{l , k}d_l\sum_{n , n'}\sum_{n''}r^{j , l}_{n , n',k}\left(\frac{r_+}{r_-}\right)^{n''}(w_l^\dag)_{n'',n}(w_l)_{n',n''}|j , n+k\>\<j , n'+k|\otimes\openone_{d_j } , \end{split}\ ] ] where @xmath220 are the entries of the wigner rotation matrix in the @xmath205 representation which rotates the @xmath83-components into the @xmath211-components in the usual notation ( the one that is found , for example , in @xcite ) such entries are denoted as @xmath221 . as discussed previously in sect .
[ s : symm - broad ] , the projection @xmath93 along the @xmath211 axis ( [ eq : rprime - phase ] ) of the bloch vector of the single - site output state is a convex ( linear ) function on the convex set of phase covariant broadcasting maps , and therefore it achieves its maximum on extremal broadcasting maps .
let the functions @xmath132 and @xmath222 denote an extremal map .
hence , starting from eq .
( [ eq : rprime - phase ] ) and specializing eq .
( [ eq : phase - form - pr ] ) , derived in appendix b , to the extremal case @xmath132 and @xmath222 , we can express the scaling factor in the following form @xmath223_{n , n+1}\left[j_x^{(j_l)}\right]_{n+k_l , n+k_l+1},\ ] ] where @xmath224_{n , m}$ ] denotes the matrix element of the operator @xmath225 evaluated with respect to the eigenstates of @xmath199 , i. e. @xmath84 .
the final form in eq .
( [ rprime - to - opt ] ) is now suitable to be optimized .
first of all , since the matrix elements @xmath226_{n , m}$ ] are non - negative , the maximum purity is reached by maximizing the off - diagonal elements of @xmath63 , namely for rank - one @xmath63 with all the matrix elements equal to @xmath227 ( see eq .
( [ eq - mancante ] ) ) .
we now want to identify the values of @xmath135 and @xmath228 corresponding to the optimal scaling factor of the map .
the matrix elements of @xmath229 take the explicit form @xmath230_{n+k_l , n+k_l+1}=\frac{1}{2}\sqrt{j_l(j_l+1)-(n+k_l ) ( n+k_l+1)}\;.\ ] ] since the above matrix elements are maximized in the central block of the matrix , the optimal map is achieved by choosing @xmath228 as close as possible to zero , for all the values of @xmath231 .
when @xmath207 is even , the optimal choice corresponds to @xmath232 for all @xmath231 .
when @xmath207 is odd there are two equivalent possible choices for @xmath228 , namely @xmath233 , for each value of @xmath231 .
moreover , we set @xmath161 for all @xmath231 , namely as large as possible . for @xmath207
even , the global output @xmath103 and the scaling factor are given by @xmath234_{n , n'}\left|\frac{m}{2},n\right\>\left\<\frac{m}{2},n'\right|,\\ & p^{n , m}_e(r)=\frac{4}{mr}(r_+r_-)^{n/2}\sum_{l = l_0}^{n/2}d_l\sum_{n =- l}^{l-1}\left[\exp\left(j_x^{(l)}\log\frac{1+r}{1-r}\right)\right]_{n , n+1}\left[j_x^{(m/2)}\right]_{n , n+1}. \end{split}\ ] ] for @xmath207 odd we have many more solutions , corresponding to all the possible combinations of @xmath235 for all values of @xmath231 .
as will be clear in the following discussion , we will examine the two cases of @xmath236 and @xmath237 for all @xmath231 . in the former case we can write @xmath238_{n , n ' } \left|\frac m2,n+\frac 12\right\>\left\<\frac m2,n'+\frac 12\right|,\\ & p^{n , m}_o(r)=\frac{4}{mr}(r_+r_-)^{n/2}\sum_{l = l_0}^{n/2}d_l\sum_{n =- l}^{l-1}\left[\exp\left(j_x^{(l)}\log\frac{1+r}{1-r}\right)\right]_{n , n+1}\left[j_x^{(m/2)}\right]_{n+1/2,n+3/2 } , \end{split}\ ] ] while for @xmath237 we have @xmath239_{n , n ' } \left|\frac m2,n-\frac 12\right\>\left\<\frac m2,n'-\frac 12\right|,\\ & p^{n , m}_o(r)=\frac{4}{mr}(r_+r_-)^{n/2}\sum_{l = l_0}^{n/2}d_l \sum_{n =- l}^{l-1}\left[\exp\left(j_x^{(l)}\log\frac{1+r}{1-r}\right ) \right]_{n , n+1}\left[j_x^{(m/2)}\right]_{n-1/2,n+1/2}. \end{split}\ ] ] notice that , since @xmath226_{m , m+1}= \left[j_x^{(j)}\right]_{-(m+1),-m}$ ] and the same property holds for the matrix elements of any power of @xmath240 , the scaling factors corresponding to the extremal maps with @xmath236 and @xmath237 are exactly the same .
this means that the bloch vector components in the @xmath192-plane are scaled in the same way by the two maps .
we want to point out that an extremal map with @xmath241 generates output density operators with a non vanishing component of the bloch vector along the @xmath83 direction .
actually , for the input state ( [ decpc ] ) the output single - site density operator is given by @xmath242 where @xmath243 optimal broadcasting maps , where the bloch vector is just scaled along its input direction , can then be obtained for odd values of @xmath207 by equally mixing the two maps considered above , corresponding to @xmath236 and @xmath237 . as mentioned earlier , since the two maps give the same scaling factor , their mixture does not compromise optimality .
notice that the optimal broadcasting maps we have derived in this way are independent of the input state . in the limit of pure input states ,
the above maps coincide with the optimal phase covariant cloning for pure equatorial states presented in ref .
@xcite .
we will now discuss more quantitatively the results derived above .
the optimal scaling factors , reported in eqs .
( [ eq : phase - pr1 ] ) and ( [ eq : phase - pr2 ] ) , contain only known terms and can be studied numerically .
versus @xmath82 for the phase covariant broadcasting , for @xmath167 and @xmath3 ranging from 4 to @xmath244 in steps of 8.,width=340 ] it turns out that phase covariant superbroadcasting is possible even for @xmath5 , with @xmath245 .
moreover , it is possible to superbroadcast an infinite number of output copies starting from @xmath4 ( @xmath183 for @xmath246 ) . as for the universal case
, we can easily compute the function @xmath141 for @xmath167 and @xmath168 , which is monotone decreasing in @xmath13 . in fig .
[ pr - n+1 ] we report the plots of @xmath170 for values of @xmath3 such that @xmath247 in steps of 8 , and for @xmath167 . in fig .
[ rstarphase ] we report the plots of the values of @xmath184 and @xmath185 , as defined in the universal case . and @xmath185 in the phase covariant broadcasting , for @xmath248 .
the upper line , corresponding to @xmath191 , has a power law @xmath249 .
the lower line , corresponding to @xmath167 , has a power law @xmath250.,width=340 ] the upper line refers to the case @xmath251 and shows a behaviour like @xmath249 .
the lower line is for @xmath252 and scales like @xmath253 . as before , in the limit of pure input states ( @xmath189 ) , the optimal phase covariant superbroadcasting map coincides with the optimal phase covariant cloner for qubits of ref .
in this section we analyze the entanglement properties of the output state of the optimal @xmath0 broadcasting maps .
notice that , since broadcasting maps are always optimized with @xmath161 , the output state is supported on @xmath254 ( which has multiplicity @xmath255 ) , namely the completely symmetric subspace of @xmath256 . therefore , also the reduced state of two qubits @xmath257 $ ] is symmetric .
we will analyze in particular bipartite entanglement in the output state , which is conveniently described in terms of the concurrence @xcite @xmath258 where @xmath259 are the decreasingly - ordered eigenvalues of the operator @xmath260 , and @xmath261 .
we will first consider the universal case , where the output state @xmath103 is diagonal on the @xmath85 basis . as shown in appendix [ appendixa ] , the state @xmath262 commutes with @xmath263 , and therefore it can be written as a linear combination of independent powers of @xmath264 , namely @xmath265 in the above expression the positivity and unit trace constraints are given by @xmath266 the eigenvalue @xmath267 of @xmath268 is always 0 , corresponding to the null component of @xmath262 on the singlet . by the unit trace condition we can express @xmath269 as a function of @xmath270 and @xmath154 as @xmath271 , and the positivity condition in terms of the two independent parameters @xmath272 is just @xmath273 the above inequality defines a triangle with basis @xmath274 $ ] and height @xmath275 $ ] , as shown in fig . [ concurr ] ( left ) .
a state in @xmath276 is then completely determined by the couple @xmath272 .
notice that the only pure states of the form ( [ eq : rho2 ] ) are @xmath277 , @xmath278 and @xmath279 , which correspond to the vertices @xmath280 , @xmath281 and @xmath282 respectively of the triangle in fig .
[ concurr ]
. we will now express the concurrence in terms of @xmath272 .
since @xmath262 is real it follows that @xmath283 , and therefore we can write @xmath284 it is easy to verify from eq .
( [ eq : rho2 ] ) that @xmath285 corresponds to the couple @xmath286 .
moreover , since @xmath262 and @xmath285 commute , the operator @xmath268 can be simply written as @xmath287 . by exploiting some algebra , and taking into account the identities @xmath288 and @xmath289
, we get the following expression @xmath290 > from the above expression , by using the unit trace constraint , we can compute the eigenvalues of @xmath268 @xmath291 notice that the first eigenvalue is doubly degenerate .
the concurrence can then be written as follows @xmath292 the above equation defines a parabola inside the triangle ( [ eq : triangle ] ) of states .
such a parabola separates the region of separable states from that of entangled states , shown in light and dark gray in fig .
[ concurr ] respectively . in order to analyze the amount of bipartite entanglement in the broadcast states
, we have then to evaluate the couple @xmath272 for the reduced state of two output copies and then determine in which region of the triangle it lies . using eq .
( [ alfabeto ] ) derived in appendix c , we can numerically evaluate @xmath272 for the universal double - site reduced output density operator @xmath257 $ ] . in fig .
[ concurr ] we report the parametric plot for the case of 4 input and 5 output copies .
as we can see , the black line moves towards positive @xmath154 as the bloch vector length @xmath82 of the input state goes from 0 to 1 .
it is possible to see in the magnified plot on the right that , as @xmath82 gets close to 1 , i. e. the input state gets pure , the output exhibits bipartite entanglement , since it crosses the parabola . in the limit of pure input states , these results agree with the ones derived in ref .
@xcite . in the phase covariant case
it is not possible to carry on the same analysis , since , as we notice in appendix [ appendixa ] , the global output state does not commute with @xmath240 .
however , using the partial traces in eqs .
( [ alfabeto ] ) , ( [ alfabeto2 ] ) , and ( [ alfabeto3 ] ) , it is still possible to evaluate the concurrence numerically . in fig .
[ concuphase ] we report the plots of the entanglement @xmath293 defined in ref .
@xcite as follows @xmath294 for @xmath295 and @xmath167 as a function of the input bloch vector length @xmath82 , both in the universal and phase covariant case .
notice that , contrarily to what happens in the universal case , in the phase covariant case bipartite entanglement vanishes in the limit of pure input states .
the absolute value of @xmath296 goes to zero for increasing number @xmath3 of input copies .
in this paper we studied symmetric broadcasting maps , where @xmath3 input qubits initially prepared in the same mixed state are transformed into an output state of @xmath13 qubits , all described by the same density operator .
we considered covariant maps and we investigated in particular the universally covariant case and the phase covariant case .
we have shown that for sufficiently mixed initial states and for @xmath297 it is also possible to partially purify the single qubit output density operator in the broadcasting operation . such a new process was named superbroadcasting . the new superbroadcasting channels open numerous interesting theoretical problems .
the first problem is to extend the map to any dimension @xmath7 and to different covariance groups .
indeed , for special cases it is easy to see that increasing the dimensionality and/or reducing the set of states to be broadcast makes superbroadcasting possible with smaller @xmath3 , and even with @xmath6 input states . as a matter of fact , this is the case of universally covariant superbroadcasting from @xmath4 to @xmath298 , which can also be regarded as superbroadcasting for @xmath299 for special states of the form @xmath300 , and for the covariance group @xmath301 .
the case of dimension @xmath299 is most interesting , since it can be exploited to improve entanglement for bipartite states of qubits .
also the infinite dimensional case ( the so - called `` continuous variables '' ) turns out to be interesting , and easily feasible experimentally @xcite .
it should be emphasized that for dimension @xmath302 there are many ways of increasing purity , and certainly the most interesting case is the purification along the mixing direction of a noisy channel ( notice that most channels do not correlate , whence the produced state is the tensor product of identical mixed states ) .
another major problem is the analysis of the detrimental correlations between two outputs , e. g. to establish whether they are quantum or classical .
these correlations are exotic , in the sense that instead of increasing the local mixing as usual , they reduce it .
such mechanism is new , and deserves a more thorough analysis .
an interesting issue , for example , is that they can not be erased leaving the local state unchanged ( the de - correlating map which sends a state to the tensor product of its partial traces is non linear ) , and this raises the problem of the optimal de - correlating channel , which optimizes the fidelity between the input and the output local state .
such optimal channel can be derived using the same technique for optimal covariant maps used in the present paper . finally , distributing quantum information and in particular superbroadcasting
raises the new problem of the trade - off between broadcasting and cryptographic security . indeed , on one side , the presence of many identical uses seems to open more possibilities of eavesdropping , however the detrimental correlations may drastically reduce such possibility , and the opportunity of detecting the eavesdropping on the joint output state may be exploited to increase the security .
this work has been co - founded by the ec under the program secoqc ( contract no . ist-2003 - 506813 ) and by the italian miur through firb ( bando 2001 ) and prin 2005 .
the global density operator @xmath77 is clearly invariant under permutations of the @xmath3 qubits , and , according to the schur - weyl duality , it can be represented by the wedderburn decomposition ( [ eq : schur - weyl - dual - form ] ) @xmath303 where @xmath304 is a state on @xmath67 . in order to evaluate @xmath304
it is sufficient to evaluate the matrix elements , where @xmath84 are eigenstates of @xmath85 in the @xmath68 representation , and @xmath305 is an arbitrary state in @xmath306 . for the sake of simplicity
we can suppose that the state @xmath78 has the bloch form @xmath307 .
the problem is now to choose @xmath305 in a suitable way .
it turns out that a clever choice is given by @xmath308 where the subscript @xmath309 means that @xmath310 is a vector in the symmetric subspace of the first @xmath309 qubits , while @xmath311 is a singlet , supporting an invariant representation on a couple of qubit spaces .
notice that in eq . the tensor product on the l.h.s
. refers to the `` abstract '' subspace @xmath312 in the wedderburn decomposition , whereas the one on the r.h.s .
refers to the decomposition @xmath313 grouping separately the first @xmath309 qubits and the remaining @xmath314 .
moreover , since the chosen density operator @xmath77 commutes with the total @xmath85 , given by the following expression @xmath315 then @xmath304 commutes with @xmath199 , which implies that @xmath304 is diagonal on the eigenstates @xmath84 of @xmath199 .
therefore we can write @xmath316 since @xmath310 is symmetric , it is a linear combination of factorized vectors with @xmath317 qubits in the @xmath318 state and @xmath319 in the state @xmath320 . as a consequence we can also write @xmath321 where @xmath322 . by analogous arguments it follows that @xmath323 the matrix element @xmath324 has the following expression @xmath325 and the decomposition of @xmath77 is finally given by @xcite @xmath326 notice that this expression exhibits a singularity for @xmath189 due to the rearrangement of terms ( [ rearr ] ) .
however , a finite limit for @xmath327 exists , as it can be seen from the equivalent expression @xmath328 which exhibits no singularities .
in this appendix we will derive the explicit form of the scaling factor for the universal and phase covariant cases . in order to calculate @xmath149 we start rewriting eq .
( [ eq : to - calculate ] ) as @xmath329=\tr\left[|ln\>\<ln|\tr_j\left[\left(j_z^{(j)}\otimes\openone_{2l+1}\right)\ p^{j}_{j , l}\right]\right].\ ] ] let us define @xmath330,\ ] ] where @xmath331 , and @xmath332 .
since @xmath333 the set @xmath334 transforms according to @xmath335\\ & = \tr_{j}\left[\left(\sum_{k=-1}^1(u^{(1)}_g)_{ik}j_k^{(j)}\otimes\openone_{2l+1}\right)\ p^{j}_{j , l}\right]\\ & = \sum_{k=-1}^1(u^{(1)}_g)_{ik}x_k^{(l)},\\ \end{split}\ ] ] and we conclude that @xmath336 is an irreducible tensor set .
it can then be proved by the wigner - eckart theorem that @xmath337 , and in particular @xmath338 . from the last relation and from the identity @xmath339 where @xmath340 and @xmath341 , and @xmath342 , we have @xmath343\\ & = \sum_{i=-1}^1a_i\tr\left[\left(j_i^{(j)}\otimes j_{-i}^{(l)}\right)p^j_{j , l}\right ] , \end{split}\ ] ] by using the well known identity @xmath344 we can write the explicit form of the coefficient @xmath270 @xmath345 by using the above expression we finally have @xmath346=n\frac{2j+1}{2l+1}\beta(j , j , l).\ ] ] substituting the global output state @xmath103 given by eq .
( [ eq : form - for - sigma ] ) into eq .
( [ eq : rprime - phase ] ) , it is possible to compute the scaling factor @xmath149 as @xmath347_{n'+k , n+k}\times\\ & \sum_{n''=-l}^l(w_l)_{n',n''}\left(\frac{r_+}{r_-}\right)^{n''}(w_l^\dag)_{n'',n } , \end{split } \label{app1}\ ] ] where @xmath220 are entries of the wigner rotation matrix in the @xmath205 representation which rotates the @xmath83-components to the @xmath211-components ( in the usual notation , the one that is found , for example , in @xcite , such entries are denoted as @xmath221 ) . in eq .
( [ app1 ] ) the sum over @xmath348 gives the @xmath349-th matrix element of @xmath350 .
therefore we can write @xmath351_{n',n}\left[j_x^{(j)}\right]_{n'+k , n+k}\\ & = \frac{4}{mr}(r_+r_-)^{n/2}\sum_{j , l , k}d_jd_l\sum_{n =- l}^lr^{j , l}_{n , n+1,k}\left[\exp\left(j_x^{(l)}\log\frac{1+r}{1-r}\right)\right]_{n , n+1}\left[j_x^{(j)}\right]_{n+k , n+k+1 } , \end{split}\ ] ] where in the last equality we used the fact that @xmath352 has non - null matrix elements only on the second - diagonals , and we multiplied the second line by a factor 2 considering in the sum only one of the two second - diagonals . notice that all matrix elements in the previous equations are calculated with respect to the @xmath83-oriented basis @xmath353 and @xmath354 .
in this appendix we want to calculate the following partial trace @xmath355,\ ] ] where the operator to be partially traced acts on @xmath356 , and @xmath84 are eigenstates of @xmath199 , as usual . in order to do that , we first decompose the vector @xmath84 into its components onto @xmath357 using the clebsch - gordan coefficients @xmath358 and then trace the operator @xmath359 over @xmath360 . in this way we get @xmath361=\frac{\openone}{2}+\frac{m}{2j}\sigma_z.\ ] ] we now recall a fact related to the already mentioned schur - weyl duality , by which multiplicity spaces @xmath306 in the wedderburn decomposition ( [ eq : wedderburn - for - su2 ] ) support irreducible representations of the permutation group @xmath362 of @xmath13 qubits
. hence , for any operator @xmath363 on @xmath312 one has @xmath364\otimes\openone_{d_j}.\ ] ] for convenience , let us write @xmath365 as we already did in eq .
( [ eq : clever - choice ] ) . with this choice
, we get : @xmath366&=\tr_{m-1}\left[\frac{d_j}{m!}\sum_\sigma\pi_\sigma^m\left(|j , m\>\<j , m|\otimes|\psi^-\>\<\psi^-|^{\otimes\frac{m-2j}{2}}\right)\pi_\sigma^m\right]\\ & = \frac{d_j}{m!}\left[(m-1)!(m-2j)\frac{\openone}{2}+(m-1)!2j\left(\frac{\openone}{2}+\frac{m}{2j}\sigma_z\right)\right]\\ \end{split}\ ] ] the first term in the sum comes from excluding from trace one of the @xmath367 qubits in singlet state .
the second term in the sum comes from excluding from trace one of the @xmath309 qubits in @xmath84 state , and from eq .
( [ partial - trace ] ) . rearranging the above equation
, we get the final expression @xmath368=d_j\left(\frac{\openone}{2}+\frac{m}{m}\sigma_z\right).\ ] ] consider the global output state @xmath369 .
if it commutes with the total angular momentum component along the direction @xmath83 , for example , then it is simple to prove that also @xmath370 $ ] commutes with @xmath101 .
in fact , from eq .
( [ eq : final - partial - trace ] ) , it is simple to see that @xmath371\propto\sigma_z,\ ] ] for all @xmath68 , and consequently also @xmath372\propto\sigma_z$ ] . in the universal case ,
the global output @xmath103 is diagonal on eigenstates of @xmath373 , hence @xmath370 $ ] commutes with @xmath101 , according to previous arguments . in the phase covariant case , it is more difficult to prove on general grounds that @xmath374=0 $ ] , since in the phase covariant case @xmath375\neq 0 $ ] .
the simplest thing we can do is to compute the partial trace of eqs .
( [ eq : phase - pr1 ] ) and ( [ eq : phase - pr2 ] ) using again clebsch - gordan coefficients ( [ eq : cg - coeff ] ) .
first of all let us notice that , tracing over @xmath376 qubits , only the terms with @xmath377 contribute . among these ,
the terms with @xmath378 give one factor proportional to @xmath379 and one factor proportional to @xmath101 , whereas terms with @xmath380 contribute with factors proportional to @xmath106 , since the matrix of coefficients @xmath381 is symmetric , see eq .
( [ eq_che_manca ] ) . then , posing @xmath382 without loss of optimality , @xmath166 $ ] commutes with @xmath106 .
we will show here how to compute the reduced output state of two copies , which is used in section [ sec : concurrence ] to compute the concurrence between two of the @xmath13 clones .
clearly it does not matter which two clones we are considering , since the global output state is permutation invariant . using clebsch - gordan coefficients , it is possible to decompose a vector in @xmath67 into its components onto @xmath383 @xmath384 and then to compute the partial trace of @xmath385 over @xmath386 .
for @xmath387 we have @xmath388=\begin{pmatrix } \frac{(m-2m)(m-2m-2)}{4m(m-1)}&0&0\\ 0&\frac{m^2 - 4m^2}{2m(m-1)}&0\\ 0&0&\frac{(m+2m)(m+2m-2)}{4m(m-1 ) } \end{pmatrix},\ ] ] when @xmath389 @xmath390=\\ & \begin{pmatrix } 0&\frac{(m-2m-2)\sqrt{(m-2m)(m+2m+2)}}{2\sqrt{2}m(m-1)}&0\\ \frac{(m-2m-2)\sqrt{(m-2m)(m+2m+2)}}{2\sqrt{2}m(m-1)}&0&\frac{(m+2m)\sqrt{(m-2m)(m+2m+2)}}{2\sqrt{2}m(m-1)}\\
0&\frac{(m+2m)\sqrt{(m-2m)(m+2m+2)}}{2\sqrt{2}m(m-1)}&0 \end{pmatrix } , \end{split}\ ] ] and @xmath391 , @xmath392=\\ & \begin{pmatrix } 0&0&\frac{\sqrt{(m-2m)(m+2m+4)(m+2m+2)(m-2m-2)}}{4m(m-1)}\\ 0&0&0\\ \frac{\sqrt{(m-2m)(m+2m+4)(m+2m+2)(m-2m-2)}}{4m(m-1)}&0&0 \end{pmatrix}. \end{split}\ ] ] for @xmath393 , partial trace over @xmath394 copies gives null contribution .
99 l. grover , quant - ph/9704012 ; j i cirac , a ekert , s f huelga , and c. macchiavello , phys .
a * 59 * , 4249 ( 1999 ) .
c crepeau , d gottesman , and a smith , in proceedings of the 34th annual acm symposium on theory of computing ( new york , acm press , 2002 ) .
s c benjamin and p m hayden , phys .
a * 64 * , 030301 ( 2001 ) . w k wootters , w h zurek , nature * 299 * , 802 ( 1982 ) .
d dieks , phys .
a , * 92 * , 271 ( 1982 ) .
h p yuen , phys .
a*113 * 405 ( 1986 ) . in ref .
@xcite it was shown that the cloning machine violates the superposition principle , which applies to a minimum total number of _ three _ states , and hence does not rule out the possibility of cloning _ two _ nonorthogonal states .
it is violation of unitarity that makes cloning any _ two _ nonorthogonal states impossible , as proved in ref . @xcite . in reference
@xcite it was shown that the proposal for superluminal communication does not work , proving the impossibility of cloning in this particular context due to linearity of evolution . according to ref .
@xcite , previous to refs .
@xcite the anonymous referee s report of g. ghirardi to ref .
@xcite contained an argument which was a special case of the no - cloning theorem of refs . @xcite .
more recently , after several attempts of determining the wave function of a single system appeared in the literature , ref .
@xcite showed how it is impossible to determine the wave function from a single copy of the system , and connected such impossibility to the no - cloning theorem .
g m dariano and h p yuen , phys .
lett . * 76 * 2832 ( 1996 ) .
n herbert , found . phys .
* 12 * , 1171 ( 1982 ) . a peres , fortsch . phys . * 51 * , 458 ( 2003 ) .
h barnum , c m caves , c a fuchs , r jozsa , and b schumacher , phys .
76 * 2818 ( 1996 ) .
r clifton , j bub , and h halvorson , found . of phys .
* 33 * 1561 ( 2003 ) .
g m dariano , c macchiavello , and p perinotti , phys .
lett . * 95 * , 060503 ( 2005 ) .
r f werner , phys . rev .
a * 58 * , 1827 ( 1998 ) .
m. keyl and r. f. werner , ann . h. poincar , * 2 * , 1 ( 2001 ) .
g m dariano , p perinotti , and m f sacchi , quant - ph/0602037 ; g m dariano , p perinotti , and m f sacchi , quant - ph/0601114 .
j i cirac , a k ekert , and c macchiavello , phys .
* 82 * , 4344 - 4347 ( 1999 ) .
f buscemi , g m dariano , c macchiavello , and p perinotti , in proceedings of the 13th quantum information technology symposium ( qit13 ) ( sendai , japan , 2005 ) .
f buscemi , g m dariano , and m f sacchi , phys .
a * 68 * , 042113 ( 2003 ) .
a jamiokowski , rep .
phys . * 3 * , 275 ( 1972 ) ; m - d choi , lin .
alg . appl .
* 10 * , 285 ( 1975 ) .
w fulton and j harris , _ representation theory : a first course _ ( springer - verlag , berlin , 1991 ) .
a r edmonds , _ angular momentum in quantum mechanics _ ( princeton university press , princeton , 1960 ) . v buek , m hillery , and r f werner , phys .
a * 60 * , r2626 ( 1999 ) .
g m dariano and c macchiavello , phys .
a * 67 * , 042306 ( 2003 )
. f buscemi , g m dariano , and c macchiavello , phys .
rev a * 72 * , 062311 ( 2005 ) . w k wootters , phys . rev .
* 80 * , 2245 ( 1998 ) .
d bru and c macchiavello , found .
phys . * 33 * ( 11 ) , 1617 ( 2003 ) . | we describe a general framework to study covariant symmetric broadcasting maps for mixed qubit states .
we explicitly derive the optimal @xmath0 superbroadcasting maps , achieving optimal purification of the single - site output copy , in both the universal and the phase covariant cases .
we also study the bipartite entanglement properties of the superbroadcast states . |
quantitative determination of the shape of capillary surfaces has been one of the standard problems in hydrodynamics , and several different techniques have been developed to deal with this .
one typical example is the stationary pending droplet . from its surface profile
one can deduce the surface tension @xcite . nowadays
, digital imaging is used to determine the edges of the photographed object , and the intensity change at the surface is imaged with a limited amount of pixels on the camera chip . a standard high resolution procedure used to reconstruct the interface
is based on an evaluation of the intensity gradient @xcite by using a gauss filter and a sobel operator @xcite .
the latter is a discrete differentiation operator that calculates the gradient of the image intensity function .
typically , heuristic fitting functions , e.g. a sigmoid , are then used to model the interface , and the resolution can be improved by an educated guess on the model function . in this way
, one goes already beyond the ideal view that the imaged contours correspond to a step function , and these procedures allow significant sub - pixel resolution fold objective under white light illumination .
left : measured diffracted light intensity pattern from a typical section of the right image .
the straight red lines indicate the intensity level at infinity and at @xmath2 of this level .
the one - fourth level indicates the actual position of the interface .
the inset shows the diffraction pattern for the case of a fully absorbing object of diameter @xmath3 , illuminated by monochromatic light of wavelength @xmath4 ; @xmath5 is the distance perpendicular to the optical axis and q is related to the edge - detector distance ( see text ) . ]
there are different approaches to improving the optical resolution beyond abbe s diffraction limit @xcite , often by evaluating the intensity profile of the diffracted light in the near field .
a simple fitting function can be used to evaluate the shape of the interface .
this super - resolution technique was being developed several years ago @xcite and applied successfully in near field scanning optical microscopy at a resolution down to 12 nm@xcite . in ref .
@xcite some of the authors of the present studies have shown that in extensional rheometry experiments one can resolve the amplitude of the sinusoidal deformation of the cylindrical liquid filament down to a resolution of @xmath6 nm . a capillary break - up rheometer ( caber ) consists of two steel plates with a droplet of a viscoelastic sample liquid placed between them .
the steel plates are drawn apart and due to the action of the capillary forces , the capillary bridge starts to shrink , and , at least in polymer solutions , one typically observes a long lasting filament that shrinks exponentially in time and eventually becomes unstable to a rayleigh - plateau like instability .
the flow in the filament is purely elongational and the polymers are very efficiently stretched , therefore , the resistance to the flow is much stronger than in shear flow , which is described by the elongational viscosity . with knowledge of the surface tension one can deduce this elongational viscosity by simply evaluating the thickness of the filament as a function of time . in simple , newtonian fluids ,
the elongational viscosity is simply three times the shear viscosity due to geometrical scaling ; this is trouton s ratio .
the aim of the present manuscript is to describe the optical procedure in more detail .
our experimental method is closely related to what is called laser diffraction edge detection in industrial applications @xcite .
the edge position of an object from its fresnel diffraction pattern can be detected either by illumination with a monochromatic light of wavelength @xmath4 or with a polychromatic light of a given bandwidth @xmath7@xmath4 .
the diffraction pattern in the near field depends on the distance of the object from the source , and the full relation between the object and the electromagnetic field in the full space is quite complex .
however , if one is interested only in the position of the edge , one can devise a simple formula from which information on the position of the object s edge can be deduced , and where only the ( averaged ) distance between the imaged electromagnetic intensity distribution and the object has to be taken as a fitting parameter .
the inset of fig .
[ fresnelbeugung ] shows the textbook case of fresnel diffraction with monochromatic light at the edge of a fully absorbing object , as well as the distances of the intensity maxima and minima from the edge scale with 1/@xmath8 , where q is the distance between the object and ( imaged ) measuring position .
it can be shown that the edge of the object is located at @xmath2 of its intensity at infinity @xmath9 @xcite . in fig .
[ fresnelbeugung ] the measured intensity distribution pattern around one edge is shown in the case of white light interacting with a human hair .
our home - made caber consists of two steel discs with a diameter of @xmath10 .
the lower disc is fixed and the upper disc can be drawn apart with a linear motor ( p01 - 23x80 , linmot , spreitenbach , switzerland ) . a @xmath11 halogen lamp with a dichroic reflector is used to produce a shadow graph image of the capillary bridge .
optically , the fluid filament is similar to a cylindrical lens .
the light beam that passes through the center of the filament is not diffracted and produces a bright line in the middle of the filament ; the light that passes close to the edges gets diffracted and so they appear dark .
the optical set - up for the caber is shown in fig .
[ setup ] .
the filament is filmed by a @xmath12 high speed camera ( x - stream xs-5 , idt , tallahassee , usa ) at @xmath13 frame rate and @xmath14 shutter time .
the camera has @xmath15 pixels with a size of @xmath16 .
microscope objectives ( nikon ) with different magnifications of 10 and 20 fold were used .
the latter was a long working distance objective ( @xmath17=7 mm ) with a numerical aperture of @xmath18=0.45 . at the shortest light wave length with reasonable sensitivity of the camera of @xmath19
this yields a diffraction limited resolution of @xmath20=0.6@xmath21 m .
such a length is magnified to 12@xmath21 m on the ccd chip .
the depth of field was @xmath22 , which made it very difficult to obtain sharp images during the final stages .
filaments with a thickness of the order of @xmath23 near the end of the thinning process were very sensitive to the slightest distortions caused by air currents , or showed non reproducible lateral movements in the order of several microns even when air currents were mostly suppressed by an additional glass box around the capillary bridge .
calibration with a reference object is needed in order to carry out this measurement method correctly . for this
propose we use an optical glass fiber manufactured by nu - fern , with specified diameter of @xmath24 .
with @xmath25 the characteristic time scale ( red line ) .
the events between 3 and 4 correspond to the formation of the first bead at the lower end of the filament .
the first generation of beads on the filament appear between 4 and 5 ; from 5 to 6 the filament thins linearly . ]
our sample solution was a 1000 or 2000 ppm polyethylene oxide ( peo ) with a mw of @xmath26 amu dissolved in water .
its physical properties under elongation are reported in several publications @xcite .
figure [ timeline ] shows the evolution of the fluid capillary bridge .
after the filament is formed , one sees exponential shrinking until it becomes unstable to a rayleigh - plateau instability and then a final pinch off which is governed by self - similar laws . at this stage ,
the shrinking of the filament follows a linear law @xcite .
both time periods , the exponential thinning between the numbers 1 and 2 , and the linear thinning between 5 and 6 are indicated with the corresponding fitting curves ( red lines ) , and the first wavy surface distortion on the filament ( sinusoidal waves or single droplets ) appears between the numbers 3 and 4 ; the first generation of beads is visible in the photographs from 4 and 5 .
we analyze the intensity profile of diffracted light at the edge of an object , i.e. the capillary bridge .
the intensity distribution of the electromagnetic light field around the object is imaged by the microscope lens on the cmos chip . and we and we only have to evaluate the near field by use of fresnel theory @xcite by considering a white illumination
. the calculation should take into account the integration over the spectral range that is determined by the cmos chip sensitivity . in our case
, it is the range from @xmath27 to @xmath28 nm , with a peak around @xmath29 nm and still non negligible sensitivity in the near infra - red . in principle , for the light source we have to consider the differential radiant energy density of a black body @xmath30 with @xmath31 the planck s constant , @xmath32 the boltzmann s constant , @xmath33 the speed of light , @xmath34 the frequency of the light , and we estimate a temperature of @xmath35=3000k .
however , our halogen spot is equipped with a dichroic mirror and in this combination the spectrum is better approximated by the rectangular window function , which is @xmath36 around a central wave length @xmath4 within the wide band @xmath7@xmath4 and @xmath37 outside this range .
we define the two fresnel integrals @xmath38 when the midsection of the liquid filament is at a distance p from a point source , the intensity distribution at a point p(q , x ) with q being the distance from the filament along the optical axes and @xmath5 the one perpendicular to it is given by @xmath39 with @xmath40 the intensity of the incoming light and the auxiliary function @xmath41 for our case where the light source is far away from the filament compared to the near field ( @xmath42 ) it follows that : @xmath43 for the experimental data evaluation q is taken as a fit parameter . in principle one would have to integrate q over the thickness of the depth of field but our main interest lies in the steep first flank of the first fresnel peak that is not sensitive to this integration procedure and instead we can fit a kind of effective q. we still have to determine the bandwidth @xmath44 of the light source ; the set of curves for different values of @xmath44 is shown in fig .
[ wideband ] .
again we find that the rising flank of the first fresnel peak is not sensitive to @xmath44 . in the inset of fig .
[ wideband ] the theoretical curve for @xmath45 is compared to the experimental result for the profile of a human hair from fig .
[ fresnelbeugung ] .
similar agreement between the measured points and these calculations is observed for an optical fiber in fig .
[ fresnelfiber ] . .
the inset shows the measurement for a section of the human hair of fig .
[ fresnelbeugung ] by folding the two edges together ( right and left triangles ) and comparing with the calculated pattern for a band width @xmath7@xmath4= @xmath46 nm around the peak intensity wave length of @xmath47 of our cmos - chip . ]
fold objective under white light illumination . left : measured points for one of the edges of the optical fiber and calculated ( line ) intensity distribution around the optical fiber for a band width @xmath7@xmath48 nm with central wave length @xmath49 nm .
the inset shows the full width of the fiber from which the thickness can be determined . ]
as q is a fitting parameter , the technique works best if we calibrate the set - up with an object of known dimensions and geometry .
therefore , we measure an optical fiber ( nu - fern , with well known specifications ( fig .
[ fresnelfiber ] ) .
the calibration is not only to introduce the length scale into our algorithm , but also for mapping the optical aberration of the @xmath1 fold objective . along the @xmath50- axis of the fiber shown in fig .
[ fresnelfiber ] . ] for this calibration , in front of the objective we set up a hanging optical fiber with a weight glued on one end that acts as a plumb line .
this allow us to assume the fiber is a straight line , thus we are able to map the optical aberration by moving this vertical in front of the lens .
once we get this map of the aberration , we use it to correct the image .
the inset of fig .
[ glassfiber ] shows the corrected profile of the optical fiber after the identification and mapping of the objective pincushion aberration . around the average diameter of the fiber of @xmath51 , .
the distribution calculated from this profile of the fiber can be fitted with a gaussian ( fig .
[ glassfiber ] ) , where the main peak corresponds to the diameter @xmath51 , with a @xmath52=@xmath53 nm .
the measured value is in good agreement with the manufacturer s specifications @xmath24 .
bottom : measublack intensity distribution across filament sections ( @xmath5-axis ) for different distances along the filament ( @xmath50-axis ) as they are indicated in the top image at @xmath54 ( square ) , @xmath55 ( up - triangle ) , @xmath56 ( circle ) and @xmath57 ( down - triangle ) .
the straight line at @xmath2 of the intensity at infinity indicates the level used for the super - resolution algorithm to measure the thickness of the filament . ]
fold objective under white light illumination .
they show a sequence in the time period of the wavy pattern that gives origin to the first generation of beads on the liquid filament . by taking the onset of the wavy distortion at t=0 ms in ( a ) , they correspond to the times ( b ) @xmath58 ms , ( c ) @xmath59 ms , ( d ) @xmath60 ms , ( e ) @xmath61 ms , ( f ) @xmath62 ms . ] ( a to e ) . ]
in this section we report the performance of our super - resolution technique by measuring the thinning capillary bridge of a droplet of @xmath63 ppm peo . of particular interest
is the time period involving the formation of the first droplets or beads on the filament ; this period was indicated in fig .
[ timeline ] by the range from 3 to 4 .
[ fresnelprofile ] shows a typical example of the intensity diffraction patterns for different sections across the mentioned liquid bridge or filament .
the physical mechanism of the capillary break - up of this solution is described in detail in ref . @xcite while here we concentrate on the application of this edge - detection technique .
our algorithm evaluates the fresnel diffraction pattern in the near field as described above . from the intensity plateaus on both sides of the border it is possible to identify the intensity level at infinity @xmath9 .
the edges of the filament are situated at the intensity level @xmath64 .
the time period of the formation of the first droplets is shown in the sequence of fig .
[ sequence ] .
a wavy distortion is the beginning of the formation of beads on the filament . in the first photographs in fig .
[ sequence ] our algorithm detects a wavy pattern with an amplitude of approximately 100 nm in this image ( fig .
[ fresnelsequence ] ( a ) ) .
this pattern grows in time and moves along the filament . in fig .
[ sequence](e ) it begins to be visible when looking at the photo with the naked eye , and it is very clear in fig . [
sequence](f ) .
the correspondence between the peaks for each profile reveals that the beads are the wavy pattern as observed in fig .
[ fresnelsequence ] ( a ) to ( e ) .
to evaluate the resolution of the super - resolution algorithm , we fit the profiles of the photographs ( a ) and ( b ) of the sequence of the fig .
[ sequence ] with a sum of sinus , eq .
( [ sumsin ] ) ( see fig .
[ hundertnanometers ] top ) .
@xmath65 we perform the frequency analysis after subtracting this fitting function from the data and we obtain the corresponding distribution for the noise ( see fig .
[ hundertnanometers ] bottom ) .
this results in a gaussian distribution with a @xmath52=@xmath66 nm . in spite of the noise , the center of the mass of the wavy oscillation
can be resolved and described , for example , with amplitudes @xmath67 no larger than @xmath0 nm and fitting errors of @xmath68 nm . ( a ) and ( b ) and corresponding fittings with equation [ sumsin ] .
bottom : gaussian distribution that results from the diameter @xmath3 profiles . ]
this method was used by sattler et.al . in ref @xcite without a detailed description of the optical method .
we have described a new experimental method to evaluate the surface profile of a capillary interface with high precision .
the method is based on the evaluation of the wide band fresnel diffraction in the near field .
the set - up has no specific requirements provided that good microscope objectives with a reasonable numerical aperture are used .
we have used this method to characterize two model objects , a human hair and a glass fiber and we have shown a hydrodynamic application on a capillary extensional rheometer . with our technique it is possible to reach sub - pixel resolution for spatial variances with less than @xmath0 nm . | we report a technique based on fresnel diffraction with white illumination that permits the resolution of capillary surface patterns of less than @xmath0 nanometers .
we investigate rayleigh plateaux like instability on a viscoelastic capillary bridge and show that we can overcome the resolution limit of optical microscopy .
the viscoelastic filaments are approximately @xmath1 microns thick at the end of the thinning process when the instability sets in .
the wavy distortions grow exponentially in time and the pattern is resolved by an image treatment that is based on an approximation of the measured rising flank of the first fresnel peak
. example.eps |
metabolic reactions enable cellular function by converting nutrients into energy and by assembling macromolecules that sustain the cellular machinery @xcite .
cellular metabolism is usually thought of as a collection of pathways comprising enzymatic reactions associated with broad functional categories . yet
metabolic reactions are highly interconnected : enzymes convert multiple reactants into products with other metabolites acting as co - factors ; enzymes can catalyse several reactions , whereas some reactions are catalysed by multiple enzymes , and so on .
this enmeshed web of reactions is thus naturally amenable to network analysis , an approach that has been successfully applied to different aspects of cellular and molecular biology , e.g. , protein - protein interactions @xcite , transcriptional regulation @xcite , or protein structure @xcite .
tools from network theory @xcite have previously been applied to the analysis of structural properties of metabolic networks , including their degree distribution @xcite , the presence of metabolic roles @xcite , and their community structure @xcite . a central challenge
, however , is that there are multiple ways to construct a ( mathematical ) graph from a metabolic network @xcite .
for example , one can create a graph with metabolites as nodes and edges representing the reactions that transform one metabolite into another @xcite ; a graph with reactions as nodes and edges corresponding to the metabolites shared among them @xcite ; or even a bipartite graph with both reactions and metabolites as nodes @xcite .
importantly , the conclusions of graph - theoretical analyses are highly dependent on the chosen graph construction @xcite .
a key feature of metabolic reactions is the directionality of flux : metabolic networks contain both irreversible and reversible reactions , and reversible reactions can change their direction depending on the cellular and environmental contexts @xcite .
many of the existing graph constructions , however , lead to undirected graphs that disregard such directional information , which is central to metabolic function @xcite .
furthermore , current graph constructions are usually derived from the whole set of metabolic reactions in an organism , and thus correspond to a generic metabolic ` blueprint ' of the cell .
however , cells switch specific pathways on and off to sustain their energetic budget in different environments @xcite .
hence , such blueprint graphs might not capture the specific metabolic connectivity in a given environment , thus limiting their ability to provide biological insights in different growth conditions . in this paper
, we present a flux - based approach to construct metabolic graphs that encapsulate the directional flow of metabolites produced or consumed through enzymatic reactions .
the proposed graphs can be tailored to incorporate flux distributions under different environmental conditions . to introduce our approach
, we proceed in two steps .
we first define the _ probabilistic flux graph _ ( pfg ) , a weighted , directed graph with reactions as nodes , edges that represent supplier - consumer relationships between reactions , and weights given by the probability that a metabolite chosen at random is produced / consumed by the source / target reaction .
this graph can be used to carry out graph - theoretical analyses of metabolic organisation independent of cellular context or environmental conditions .
we then show that this formalism can be adapted seamlessly to construct the _ metabolic flux graph _ ( mfg ) , a directed , environment - dependent , graph with weights computed from flux balance analysis ( fba ) @xcite , the most widespread method to study genome - scale metabolic networks .
our formulation overcomes several drawbacks of current constructions of metabolic graphs .
firstly , in our flux graphs , an edge indicates that metabolites are produced by the source reaction and consumed by the target reaction , thus accounting for metabolic directionality and the natural flow of chemical mass from reactants to products .
secondly , the probabilistic flux graph discounts naturally the over - representation of pool metabolites ( e.g. , adenosine triphosphate ( atp ) , nicotinamide adenine dinucleotide ( nadh ) , protons , water , and other co - factors ) that appear in many reactions and tend to obfuscate the graph connectivity .
our construction avoids the removal of pool metabolites from the network , which can change the graph structure drastically @xcite .
finally , the metabolic flux graph incorporates additional biological information reflecting the effect of the environmental context into the graph construction .
in particular , since the weights in the mfg correspond directly to fluxes ( in units of mass per time ) , different biological scenarios can be analysed using fluxes from different fba solutions under different carbon sources and other environmental perturbations @xcite . after introducing the mathematical framework , we showcase our approach on the core model of _ escherichia coli _
metabolism @xcite . in absence of environmental context
, our analysis of the pfg reveals the importance of including directionality and appropriate edge weights in the graph to understand the modular organisation of metabolic sub - systems .
we then use fba solutions for several relevant growth conditions , and show that the structure of the mfg changes dramatically in each case , thus capturing the environment - dependent nature of metabolism .
consider a metabolic network composed of @xmath0 metabolites with concentrations @xmath1 ( @xmath2 ) that participate in @xmath3 reactions @xmath4 [ ] } \sum_{i=1}^n \beta_{ij } x_i}\label{eq : reac},\quad j=1,2,\ldots , m,\ ] ] where @xmath5 and @xmath6 are the stoichiometric coefficients of species @xmath7 in reaction @xmath8 .
each reaction takes place at a time - dependent rate @xmath9 , measured in units of concentration per time @xcite .
the dynamics of the metabolite concentrations is represented compactly by the system of differential equations @xmath10 where @xmath11 is the @xmath0-dimensional vector of metabolite concentrations , and @xmath12 is the @xmath3-dimensional vector of reaction rates .
the @xmath13 matrix @xmath14 is the stoichiometric matrix with entries @xmath15 , i.e. , the net number of @xmath16 molecules produced ( positive @xmath17 ) or consumed ( negative @xmath17 ) by the @xmath8-th reaction . to illustrate the different schemes and graphs described in this paper
, we use a toy example @xcite of a metabolic network including nutrient uptake , biosynthesis of metabolic intermediates , secretion of waste products , and biomass production ( figure [ fig : toymodel_networks]a ) . .
( b ) bipartite graph associated with the boolean stoichiometric matrix @xmath18 , and the standard reaction adjacency graph ( rag @xcite ) with adjacency matrix @xmath19 .
the undirected edges of @xmath20 indicate the number of shared metabolites among reactions .
( c ) the probabilistic flux graph ( pfg , @xmath21 ) and two metabolic flux graphs ( mfg , @xmath22 ) are constructed from the unfolded consumption and production stoichiometric matrices .
note that the reversible reaction @xmath23 is unfolded into two nodes
. the pfg is a directed graph with weights representing the probability that the source reaction produces a metabolite consumed by the target reaction .
the mfgs are constructed from two flux balance analysis solutions ( @xmath24 and @xmath25 ) obtained by optimizing a biomass objective function under different flux constraints representing different environmental or cellular contexts ( see sec . [
sec : toy_model ] for details ) .
the edges of the mfgs represent mass flow from source to target reactions , with weights in units of metabolic flux .
the computed fba solutions translate into different connectivity in the resulting mfgs.,scaledwidth=100.0% ] starting from the stoichiometry @xmath14 , there are several ways to construct a graph for a given set of metabolic reactions @xcite .
a common approach @xcite is to define the _ unipartite graph _ with reactions as nodes and the @xmath26 adjacency matrix @xmath27 where @xmath18 is the boolean version of @xmath14 ( i.e. , @xmath28 when @xmath29 and @xmath30 otherwise ) . in this
_ reaction adjacency graph _ ( rag ) , two reactions ( nodes ) are connected if they share metabolites , either as reactants or products , and self - loops represent the total number of metabolites that participate in a reaction ( fig .
[ fig : toymodel_networks]b ) .
though widely studied @xcite , the rag has known important limitations , as it overlooks key aspects of the connectivity of metabolic networks . in particular , the rag does not distinguish between forward and backward fluxes , nor does it incorporate information on the irreversibility of reactions ( @xmath31 is a symmetric matrix ) . furthermore , the structure of @xmath31 is dominated by the large number of connections introduced by pool metabolites ( e.g. , water , ions or enzymatic cofactors ) that appear in many reactions . although computational schemes have been designed to ameliorate the pool metabolite bias @xcite , their justification does not follow from biophysical considerations .
finally , the construction of the graph @xmath31 from rigid topological criteria is not easily extended to incorporate the effect of varying environmental contexts . to address the limitations of the standard reaction adjacency graph @xmath31 , given in eq . , and aiming at enhanced biophysical and biological interpretability , we propose a graph formulation that follows from a flux - based perspective . to construct our graph ,
we unfold each reaction into separate forward and reverse directions , and redefine the presence of links between reaction nodes to reflect producer - consumer relationships , i.e. , two reactions are connected if one produces a metabolite that is consumed by the other . as shown below
, this definition leads to graphs that naturally account for the reversibility of reactions and allows for a seamless integration of different biological contexts modelled through fba . inspired by matrix formulations of chemical reaction network kinetics @xcite
, we unfold the reactions into forward and backward components .
specifically , let us rewrite the reaction rate vector introduced in as : @xmath32 where @xmath33 and @xmath34 are non - negative vectors containing the forward and backward reaction rates , respectively , and @xmath35 is the @xmath3-dimensional reversibility vector with components @xmath36 if reaction @xmath37 is reversible and @xmath38 if it is irreversible .
the @xmath39 matrix @xmath40 contains @xmath35 in its main diagonal . with these definitions , we rewrite the metabolic model in eq . in terms of the unfolded @xmath41-dimensional vector of reaction rates , @xmath42^t$ ] , to obtain : @xmath43 where @xmath44 is the @xmath39 identity matrix , and @xmath45 is an unfolded version of the stoichiometric matrix corresponding to the @xmath41 forward and reverse reactions .
the unfolding into forward and backward fluxes leads us to the definition of _ production _ and _ consumption _ stoichiometric matrices : @xmath46 where @xmath47 is the matrix whose entries are the absolute values of the corresponding entries of @xmath45 .
note that each entry of the matrix @xmath48 , denoted @xmath49 , gives the number of molecules of metabolite @xmath16 produced by reaction @xmath50 .
conversely , the entries of @xmath51 , denoted @xmath52 , correspond to the number of molecules of metabolite @xmath16 consumed by reaction @xmath50 . within our directional flux framework ,
it is natural to consider a probabilistic description of producer - consumer relationships between reactions , as follows . given a stoichiometric matrix @xmath14 , and in the absence of further biological information , the probability that metabolite @xmath53 is produced by reaction @xmath54 and consumed by reaction @xmath37 is : @xmath55 where @xmath56 and @xmath57 are the total number of molecules of @xmath53 produced and consumed by all reactions .
we thus define the weight of the edge between reaction nodes @xmath54 and @xmath37 as the probability that _ any _ metabolite chosen at random is produced by @xmath54 and consumed by @xmath37 . summing over all metabolites and normalizing ,
the edge weight is defined as @xmath58 these edge weights are the entries of the adjacency matrix of the @xmath59 where @xmath60 , @xmath61 , @xmath62 is a vector of ones , and @xmath63 denotes the moore - penrose pseudoinverse .
the pfg is a weighted , directed graph with a double - stochastic adjacency matrix ( @xmath64 ) .
it provides a directional blueprint of the whole metabolic model , and naturally scales the contribution of pool metabolites to flux transfer . in figure
[ fig : toymodel_networks]c we illustrate the creation of the pfg for a toy network .
note that our pfg is distinct from a directed graph directly comparable to the rag @xmath65 , and with similar shortcomings , which can be constructed from boolean versions of the production and consumption stoichiometric matrices as shown sec .
[ sec : adir_vs_d_si ] .
we now extend the idea behind the construction of the pfg to account for specific environmental context or growth conditions .
cells adjust their metabolic fluxes to respond to the availability of nutrients and environmental requirements .
flux balance analysis ( fba ) is a widely used method to predict environment - specific flux distributions .
fba computes a vector of metabolic fluxes @xmath66 that maximise a cellular objective ( e.g. , biomass , growth or atp production ) .
the fba solution is obtained assuming steady state conditions and subject to constraints that describe the availability of nutrients and other extracellular compounds @xcite .
the key basic elements of fba are briefly summarised in section [ sec : fba ] . to incorporate the biological information afforded by fba solutions into the structure of a metabolic graph
, we again define the graph edges in terms of production and consumptions fluxes . similarly to eq .
, we unfold the fba solution vector @xmath67 into forward and backward components @xmath68 whence we compute the vector of production and consumption fluxes as @xmath69 the @xmath70-th entry of @xmath71 is the flux at which metabolite @xmath53 is produced and consumed . note that production and consumption fluxes are identical because of the steady state condition ( @xmath72 in eq . ) .
we now construct the flux graph by defining the weight of the edge between reactions @xmath54 and @xmath37 as the _ total flux of metabolites produced by @xmath54 that are consumed by @xmath37_. assuming that the amount of metabolite produced by one reaction is distributed among the reactions that consume it in proportion to their flux , the flux of metabolite @xmath53 from reaction @xmath54 to @xmath37 is given by @xmath73 for example , if the total flux of metabolite @xmath74 is @xmath75 , with reaction @xmath76 producing @xmath53 at a rate @xmath77 and reaction @xmath37 consuming @xmath74 at a rate @xmath78 , then the flux of @xmath53 from @xmath54 to @xmath37 is @xmath79 .
summing eq . over all metabolites
, we then obtain the edge weight relating reactions @xmath54 and @xmath37 as @xmath80 the edge weights are collected into the adjacency matrix of the @xmath81 where @xmath82 , @xmath83 and @xmath63 denotes the matrix pseudoinverse .
the mfg is a directed graph with weights in units of @xmath84 .
self loops describe the metabolic flux of autocatalytic reactions , i.e. , those in which products are also reactants .
the mfg provides a versatile framework to create environment - specific metabolic graphs from fba solutions . in figure
[ fig : toymodel_networks]c we illustrate the creation of mfgs for a toy network . in each case
we compute fba solutions under a fixed uptake flux and constrain the remaining fluxes to account for different biological scenarios . in scenario 1
the fluxes are constrained to be strictly positive and no larger than the nutrient uptake flux , while in scenario 2 we impose a positive lower bound on reaction @xmath85 .
the graph in scenario 2 displays an extra edge between reactions @xmath23 and @xmath85 and distinct edge weights , as compared to scenario 1 ( see sec . [
sec : toy_model ] ) .
the results thus illustrate how changes in the fba solutions translate into different graph connectivities and edge weights .
, as given by eq . .
the nodes represent reactions ; two reactions are linked by an undirected edge if they share reactants or products .
the nodes are coloured according to their pagerank score , a measure of their centrality in the graph .
( c ) the directed probabilistic flux graph @xmath21 , as computed from eq . .
the reversible reactions are unfolded into two overlapping nodes ( one for the forward reaction , one for the backward ) .
the directed links indicate flow of metabolites produced by the source node and consumed by the target node .
the nodes are coloured according to their pagerank score .
( d ) comparison of pagerank percentiles of reactions in @xmath20 and @xmath21 .
reversible reactions are represented by two triangles connected by a line ; both share the same pagerank in @xmath20 , but each has its own pagerank in @xmath21 .
reactions that appear above ( below ) the diagonal have increased ( decreased ) pagerank in @xmath21 as compared to @xmath20 . ] to illustrate our framework , we construct and analyse flux graphs ( pfg and mfgs ) of the core metabolic model of _ e. coli
_ @xcite .
this model ( fig .
[ fig : model_a_dnorm]a ) contains 72 metabolites and 95 reactions , grouped into 11 pathways , which describe the main biochemical routes in central carbon metabolism @xcite . to highlight the effect of flux directionality on the constructed graphs , figure [ fig : model_a_dnorm ]
compares the standard undirected reaction adjacency graph ( @xmath31 ) and our proposed probabilistic flux graph ( @xmath21 ) for the same metabolic model .
the @xmath31 graph has 95 nodes and 1,158 undirected edges , while the @xmath21 graph has 154 nodes and 1,604 directed and weighted edges . the increase in node count is due to the unfolding of forward and backward reactions into separate nodes . unlike the @xmath20 graph ,
where the edges represent shared metabolites between two reactions , the directed edges of the @xmath21 graph represent the flow of metabolites from a source to a target reaction .
a salient feature of both graphs is their high connectivity , which is not apparent from the traditional pathway representation in figure [ fig : model_a_dnorm]a .
the impact of directionality becomes apparent when comparing the importance of reaction nodes on the overall connectivity of each graph , as measured by the pagerank score introduced in the original google algorithm @xcite .
figure [ fig : model_a_dnorm]b d shows that the pagerank of reactions is substantially different in @xmath20 and @xmath21 .
the overall ordering is maintained : exchange reactions tend to have low pagerank , whereas core metabolic reactions have high pageranks in both graphs indeed , the biomass reaction has the highest rank in both cases .
however , we observe a dramatic change in the importance of most reactions .
for example , the reactions for atp maintenance ( atpm ) , phosphoenolpyruvate synthase ( pps ) and abc - mediated transport of l - glutamine ( glnabc ) drop from being among the top 10% most important reactions in the @xmath20 graph to the bottom percentiles in the @xmath21 graph .
conversely , other reactions such as aconitase a ( aconta ) , transaldolase ( tala ) and succinyl - coa synthetase ( sucoas ) , and formate transport via diffusion ( forti ) gain substantial importance in the @xmath21 graph .
for instance , forti is the sole consumer of formate , which is produced by pyruvate formate lyase ( pfl ) , a reaction that is highly connected to the rest of the network .
importantly , in most of the reversible reactions , such as atp synthase ( atps4r ) , there is a wide gap between the pagerank of the forward and backward reactions , suggesting a marked asymmetry in the importance of metabolic flows . and
@xmath21 from fig .
[ fig : model_a_dnorm]b c partitioned into communities computed with the markov stability method ; for clarity , the graph edges are not shown .
the sankey diagrams @xcite show the correspondence between biochemical pathways and the communities found in each graph .
the word clouds contain the metabolites that participate in the reactions each community , and the word size is proportional to the number of reactions in which each metabolite participates . ]
community detection is a technique that is frequently used for the analysis of complex graphs : nodes are clustered into tightly - related communities in order to reveal the coarse - grained structure of the graph , potentially at different levels of resolution @xcite .
indeed , the community structure of graphs derived from metabolic networks has been the subject of several analyses @xcite .
however , most existing community detection methods are only applicable to undirected graphs and fail to capture the directionality of the edges , a key feature in metabolism . in order to account for directionality
, we use the markov stability community detection framework @xcite , which employs diffusions on graphs to detect groups of nodes where flows are retained persistently across time scales . due to its use of diffusive dynamics , markov stability is ideally suited to find multi - resolution community structure @xcite , and naturally incorporates edge directionality , if present @xcite ( see sec . [
sec : markov ] ) . when applied to metabolic graphs
, markov stability can thus reveal groups of reactions that are tightly linked by the flow of metabolites they produce and consume .
figure [ fig : avsd_comms ] highlights the strong differences between the community structure of the undirected rag and the directed pfg of the core metabolism of _
e. coli _ , underscoring the importance of directionality in these graphs .
when applied to the @xmath20 graph , markov stability reveals a robust partition into seven communities ( figure [ fig : avsd_comms]b , see also sec .
[ sec : avsd_comms_si ] ) .
the reaction communities obtained are largely determined by the edges created by abundant pool metabolites .
for example , community c1@xmath86 is mainly composed of reactions that consume or produce atp and water .
note , however , that the biomass reaction ( the largest consumer of atp ) is not a member of c1@xmath86 because , in the @xmath20 graph construction , any connection involving atp has equal weight .
other communities in @xmath20 are also determined by pool metabolites , e.g. c2@xmath86 is dominated by h@xmath87 , and c3@xmath86 is dominated by nad@xmath87 and nadp@xmath87 , as shown by the word clouds representing the relative frequency of metabolites that appear in the reactions contained in each community .
the community structure in @xmath31 thus reflects the limitations of this graph construction due to the absence of biological context and the large number of uninformative links introduced by pool metabolites .
in contrast , we found a robust partition into five communities for the @xmath21 graph ( figure [ fig : avsd_comms]c , see also sec .
[ sec : avsd_comms_si ] ) .
these communities comprise reactions related consistently by biochemical pathways .
community c1@xmath88 contains the reactions in the pentose phosphate pathway together with the first steps of glycolysis involving d - fructose , d - glucose , or d - ribulose .
community c2@xmath88 contains the main reactions that produce atp from substrate level as well as oxidative phosphorylation and the biomass reaction .
community c3@xmath88 includes the core of the citric acid cycle , anaplerotic reactions related to malate syntheses , as well as the intake of cofactors such as co@xmath89 .
community c4@xmath88 contains reactions that are secondary sources of carbon ( such as malate and succinate ) , as well as oxidative phosphorilation reactions .
finally , community c5@xmath88 contains reactions that are part of the pyruvate metabolism subsystem , as well as transport reactions for the most common secondary carbon metabolites such as lactate , formate , acetaldehyde and ethanol .
altogether , the communities of the @xmath21 graph reflect metabolite flows associated with specific cellular functions , a key benefit and consequence of including flux directionality in the graph construction . as seen in fig .
[ fig : avsd_comms]c , the communities are no longer exclusively determined by pool metabolites ( e.g. , water is no longer dominant and protons are spread among all communities ) . for a more detailed explanation and comparison of the communities found in the @xmath20 and @xmath21 graphs ,
see section [ sec : avsd_comms_si ] .
to incorporate the impact of environmental context in our graphs , we construct different metabolic flux graphs , using flux distributions obtained from flux balance analysis applied to the core model of _
metabolism under several growth conditions : aerobic growth in rich media with glucose or ethanol , aerobic growth in glucose but phosphate- and ammonium - limited , and anaerobic growth in glucose .
the results , summarised in figure [ fig : networksm ] , reveal how changes in metabolite flows induced by different biological contexts are reflected in our graph construction . in all cases ,
the mfgs have fewer nodes than the blueprint graph @xmath21 because the fba solutions contain numerous reactions with zero flux .
the different environments also affect the graph connectivity and the relative node importance , as measured by their pagerank score .
furthermore , the community structure of the mfgs for the four environmental conditions , as obtained with the markov stability framework , reflect the distinct usage of functional pathways by the cell in response to growth requirements under specific environments .
we briefly describe the salient features of the analysis ; a more detailed discussion can be found in section [ sec : m_comms_si ] and fig .
[ fig : markov_stability_fba_networks ] in the si .
[ [ aerobic - growth - in - d - glucose - mathbfm_mathrmglc . ] ] aerobic growth in d - glucose ( @xmath90 ) .
+ + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + we observe a robust partition into three communities with an intuitive biological interpretation ( fig .
[ fig : networksm]a and fig .
[ fig : markov_stability_fba_networks]a ) .
firstly , c1@xmath91 can be thought of as a carbon - processing community , comprising reactions that process carbon from d - glucose to pyruvate including most of the glycolysis and pentose phosphate pathways , together with related transport and exchange reactions .
secondly , c2@xmath91 harbours the bulk of reactions related to oxidative phosphorylation and the production of energy in the cell , including the electron transport chain of nadh dehydrogenase , cytochrome oxidase , and atp synthase , as well as transport reactions for phosphate and oxygen intake and proton balance .
the growth reaction is also included in community c2@xmath91 , consistent with atp being the main substrate for both the atp maintenance ( atpm ) requirement and the biomass reaction in this biological scenario .
finally , c3@xmath91 contains reactions related to the citric acid cycle ( tac ) and the production of nadh and nadph ( i.e. , the cell s reductive power ) , as well as routes that take phosphoenolpyruvic acid ( pep ) as a starting point , thus highlighting carbon intake routes that are strongly linked to the tca cycle .
[ [ aerobic - growth - in - ethanol - mathbfm_mathrmetoh . ] ] aerobic growth in ethanol ( @xmath92 ) .
+ + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + we found a robust partition into three communities that resemble those found in @xmath90 , with subtle yet important differences ( fig .
[ fig : networksm]b and fig .
[ fig : markov_stability_fba_networks]b ) .
the most salient differences are observed in the carbon - processing community c1@xmath93 , which clearly reflects the switch of carbon source from d - glucose to ethanol .
this community contains gluconeogenic reactions ( instead of glycolytic ) , due to the reversal of flux induced by the change of carbon source , as well as anaplerotic reactions and reactions related to glutamate metabolism .
the main role of the reactions in this community is the production of bioprecursors such as pep , pyruvate , 3-phospho - d - glycerate ( 3pg ) , glyceraldehyde-3-phosphate ( g3p ) , d - fructose-6-phosphate ( f6p ) , and d - glucose-6-phosphate , all of which are substrates for growth .
consequently , the biomass reaction is also grouped within c1@xmath93 due the increased metabolic flux of precursors relative to atp production in this biological scenario .
the other two reaction communities ( energy - generation c2@xmath93 and citric acid cycle c3@xmath93 ) display less prominent differences relative to the @xmath90 graph , with additional pyruvate metabolism and anaplerotic reactions as well as subtle ascriptions of reactions involved in nadh / nadph balance and the source for acetyl - coa . [ [ anaerobic - growth - in - d - glucose - mathbfm_mathrmanaero . ] ] anaerobic growth in d - glucose ( @xmath94 ) .
+ + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + the absence of oxygen has a profound impact on the metabolic balance of the cell and the mfg captures the drastic changes in this new regime effectively ( fig .
[ fig : networksm]c and fig .
[ fig : markov_stability_fba_networks]c ) . both the connectivity and
the reaction communities in this mfg are different from the aerobic scenarios , with a much diminished presence of oxidative phosphorylation pathways and the absence of the first two steps of the electron transport chain ( cytbd and nadh16 ) .
we found that @xmath94 has a robust partition into four communities .
c1@xmath95 still contains carbon processing ( glucose intake and glycolysis ) , yet these reactions are decoupled from the pentose phosphate pathway , which is now part of community c3@xmath95 grouped with the citric acid cycle ( now incomplete ) and the biomass reaction .
c3@xmath95 includes the growth precursors in this scenario , including alpha - d - ribose-5-phosphate ( r5p ) , d - erythrose-4-phosphate ( e4p ) , 2-oxalacetate and nadph .
the other two communities are specific to the anaerobic context : c2@xmath95 contains the conversion of pep into formate ( more than half of the carbon secreted by the cell becomes formate @xcite ) ; and c4@xmath95 includes nadh production and consumption via reactions linked to glyceraldehyde-3-phosphate dehydrogenase ( gapd ) .
[ [ aerobic - growth - in - d - glucose - but - limited - phosphate - and - ammonium - mathbfm_mathrmlim . ] ] aerobic growth in d - glucose but limited phosphate and ammonium ( @xmath96 ) .
+ + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + under these growth - limiting conditions , we found a robust partition into three communities ( fig .
[ fig : networksm]d and fig .
[ fig : markov_stability_fba_networks]d ) .
the community structure reflects _ overflow metabolism _
@xcite , which occurs when the cell takes in more carbon than it can process . as a consequence
, the excess carbon is secreted from the cell , leading to a strong decrease in growth and a partial shutdown of the citric acid cycle .
this is reflected in the reduced weight of the tca pathway in c3@xmath97 and its grouping with the secretion routes of acetate and formate .
hence , c3@xmath97 comprises reactions that would not be strongly coupled in more favorable growth conditions , yet are linked together by metabolic responses appearing due to the limited availability of ammonium and phosphate .
furthermore , the carbon - processing community c1@xmath97 contains the glycolytic pathway , but detached from the pentose phosphate pathway , as in the @xmath94 graph , highlighting its role in precursor formation .
the bioenergetic machinery is contained in community c2@xmath97 , including the pentose phosphate pathway , with a smaller role for the electron transport chain ( 21.8% of the total atp as compared to 66.5% in @xmath90 ) .
another advantage of flux - based mfgs is the possibility of applying network - theoretic tools to detect natural groupings of reactions at different levels of resolution , as well as their hierarchical relationship across scales .
the markov stability framework @xcite can be used to detect multi - resolution community structure in directed graphs ( sec .
[ sec : markov ] ) , thus allowing the exploration of the modular multiscale organisation of metabolic reaction networks . ) across levels of resolution .
the top panel shows the number of communities of the optimal partition ( blue line ) and two measures of its robustness ( @xmath98 ( green line ) and @xmath99 ( colour map ) ) as a function of the markov time @xmath100 ( see text and methods section ) .
the five markov times selected correspond to robust partitions of the graph into 11 , 7 , 5 , 3 , and 2 communities , as signalled by extended low values of @xmath99 and low values ( or pronounced dips ) of @xmath98 .
the sankey diagram ( middle panel ) visualises the multiscale organisation of the communities of the flux graph across markov times , and the relationship of the communities with the biochemical pathways .
the bottom panel shows the five partitions at the selected markov times .
the partition into 3 communities corresponds to that in figure [ fig : networksm]a . ]
figure [ fig : vi_alluvialdiagrams ] illustrates this multiscale analysis on the metabolic flux graph of _
e. coli _ under aerobic growth in glucose ( @xmath90 ) . by varying the markov time @xmath100 , a parameter in the markov stability method
, we scanned the community structures at different resolutions .
our results show that , as we move from finer to coarser resolutions , the mfg can be partitioned into 11 , 7 , 5 , 3 , and 2 communities which have high robustness across markov time ( extended plateaux of optimality over @xmath100 , as shown by the low values of @xmath99 ) and are highly robust within the optimisation ensemble ( as shown by dips in @xmath98 )
. for further details , see section [ sec : markov ] and refs .
@xcite .
the sankey diagram in fig .
[ fig : vi_alluvialdiagrams ] allows us to visualise the pathway composition of the graph partitions and their relationships across different resolutions . as we decrease the resolution ( longer markov times ) , the reactions in different pathways assemble and split into different groupings , reflecting both specific relationships and general organisation principles associated with this growth condition .
a general observation is that glycolysis is grouped together with oxidative phosphorylation across most scales , underlining the fact that those two pathways function as cohesive metabolic sub - units in aerobic conditions .
in contrast , the exchange and transport pathways appear spread among multiple partitions across all resolutions .
this is expected , as these are enabling functional pathways in which reactions do not interact amongst themselves but rather feed substrates to other pathways .
other reaction groupings reflect more specific relationships .
for example , the citric acid cycle ( always linked to anaplerotic reactions ) appears as a cohesive unit across most scales , and only splits in two in the very final flux grouping , reflecting the global role of the tca cycle in linking to both glycolysis and oxidative phosphorylation . the pentose phosphate pathway , on the other hand , is split into two groups ( one linked to glutamate metabolism and another one linked to glycolysis ) across early scales , only merging into the same community towards the final groupings .
this suggests a more interconnected flux relationship of the different steps of the penthose phosphate pathway with the rest of metabolism .
figure [ fig : markov_stability_fba_networks ] contains a multiscale analyses of the communities for the other three growth scenarios .
metabolic reactions are commonly understood in terms of functional pathways that are heavily interconnected to form metabolic networks , i.e. , metabolites linked by arrows representing enzymatic reactions between them @xcite ( figs .
[ fig : model_a_dnorm ] and [ fig : networksm ] )
. however , such standard representations are not amenable to rigorous graph - theoretic analysis .
importantly , there are fundamentally different graphs that can be constructed from the metabolic reaction information depending on the chosen representation of species / interactions as nodes / edges , e.g. , reactions as nodes ; metabolites as nodes ; or both reaction and metabolites as nodes @xcite
. each one of these graphs can be directed or undirected , and with weighted links computed according to different rules .
the choices and subtleties in graph construction are crucial both to capture the relevant metabolic information and to interpret their topological properties @xcite . here , we have presented a flux - based strategy to build graphs for metabolic networks .
our graphs have reactions as nodes and directed edges representing the flux of metabolites produced by a source reaction and consumed by a target reaction .
this principle can be applied to build both ` blueprint ' graphs ( pfg ) , which summarise the probabilistic fluxes of the whole metabolism of an organism , as well as context - specific graphs ( mfgs ) , which reflect specific environmental conditions .
the blueprint probabilistic flux graph , with edge weights equal to the probability that the source / target reactions produce / consume a molecule of a metabolite chosen at random , naturally tames the over - representation of pool metabolites without the need to remove them from the graph arbitrarily , as is often done in the literature @xcite . the context - specific metabolic flux graphs ( mfgs )
incorporate the effect of the environment , as edge weights correspond to the total flux of metabolites between reactions as calculated by flux balance analysis ( fba ) .
computing fba solutions for different environments allows us to build metabolic graphs systematically for different growth media . to exemplify our approach
, we built and analysed pfg and mfgs for the core metabolism of _
e. coli_. through the analysis of topological properties and community structure of these graphs , we highlighted the importance of weighted directionality in metabolic graph construction and revealed the flux - mediated relationships between functional pathways under different environments .
in particular , the mfgs capture specific metabolic adaptations such as the glycolytic - gluconeogenic switch , overflow metabolism , and the effects of anoxia .
we note that although we have illustrated our analysis on the core metabolism of _
e. coli _ , the proposed graph construction can be readily applied to large genome - scale metabolic networks @xcite .
our flux graphs provide a systematic connection between network theory and constraint - based methods widely employed in metabolic modelling @xcite , thus opening avenues towards environment - dependent , graph - based analyses of cell metabolism .
an area of interest would be to use mfgs to study how the community structure of flux graphs across scales can help characterise metabolic conditions that maximise the efficacy of drug treatments or disease - related distortions , e.g. , cancer - related metabolic signatures @xcite .
in particular , mfgs can quantify metabolic robustness via graph statistics upon removal of reaction nodes @xcite .
the proposed graph construction framework can be extended in different directions .
the core idea behind our framework is the distinction between production and consumption fluxes , and how to encode both in the links of a graph .
this general principle can also be used to build other potentially useful graphs .
for example , two other graphs that describe relationships between reactions are : @xmath101 the competition and synergy graphs are undirected and their edge weights represent the probability that two reactions consume ( @xmath102 ) or produce ( @xmath103 ) metabolites picked at random .
there exist corresponding fba versions of the competition and synergy flux graphs , which follow from , and the definitions and .
these graphs could help reveal further relationships between metabolic reactions in the cell and will be the subject of future studies .
our approach could also be extended to include dynamic adaptations of metabolic activity , e.g. , by using dynamic extensions of fba @xcite or by incorporating static @xcite and time - varying @xcite enzyme concentrations .
of particular interest to metabolic modelling , we envision that mfgs could provide a novel route to evaluate the robustness of fba solutions @xcite by exploiting the non - uniqueness of the mfg from each fba solution in the space of graphs .
such results could enhance the interface between network science and metabolic analysis , allowing for the systematic exploration of the system - level organisation of metabolism in response to environmental constraints and disease states .
flux balance analysis ( fba ) @xcite is a widely - adopted approach to analyse metabolism and cellular growth .
fba calculates the reaction fluxes that optimise growth in specific biological contexts .
the main hypothesis behind fba is that cells adapt their metabolism to maximise growth in different biological conditions .
the conditions are encoded as constraints on the fluxes of certain reactions ; for example , constraints reactions that import nutrients and other necessary compounds from the exterior . the mathematical formulation of the fba is described in the following constrained optimisation problem : @xmath104 where @xmath14 is the stoichiometry matrix of the model , @xmath105 the vector of fluxes , @xmath106 is an indicator vector ( i.e. , @xmath107 when @xmath7 is the biomass reaction and zero everywhere else ) so that @xmath108 is the flux of the biomass reaction .
the constraint @xmath109 enforces mass - conservation at stationarity , and @xmath110 and @xmath111 are the lower and upper bounds of each reaction s flux . through these vectors
, one can encode a variety of different scenarios @xcite .
the biomass reaction represents the most widely - used flux that is optimised , although there are others can be used as well @xcite . in our simulations
, we set the individual carbon intake rate to 18.5 @xmath84 for every source available in each scenario .
we allowed oxygen intake to reach the maximum needed in to consume all the carbon except in the anaerobic condition scenario , in which the upper bound for oxygen intake was @xmath112 . in the scenario with limited phosphate and ammonium intake , the levels of nh@xmath113 and phosphate intake were fixed at @xmath114 and @xmath115 respectively ( a reduction of 50% compared to a glucose - fed aerobic scenario with no restrictions ) .
we extract the communities in each network using the markov stability community detection framework @xcite .
this framework uses diffusion processes on the network to find groups of nodes ( i.e. , communities ) that retain flows for longer than one would expect on a comparable random network ; in addition , markov stability incorporates directed flows seamlessly into the analysis @xcite .
the diffusion process we use is a continuous - time markov process on the network . from the adjacency matrix @xmath31
we construct a rate matrix for the process : @xmath116 , where @xmath117 is the diagonal matrix of out - strengths , @xmath118
. when a node has no outgoing edges then we simply let @xmath119 .
in general , a directed network will not be strongly - connected and thus a markov process on @xmath120 will not have a unique steady state . to ensure the uniqueness of the steady state we must add a _ teleportation _ component to the dynamics by which a random walker visiting a node can follow an outgoing edge with probability @xmath121 or jump ( teleport ) uniformly to any other node in the network with probability @xmath122 @xcite .
the rate matrix of a markov process with teleportation is : @xmath123{\mathbf{1}\mathbf{1}^t},\ ] ] where the @xmath124 vector @xmath125 is an indicator for dangling nodes : if node @xmath7 has no outgoing edges then @xmath126 , and @xmath127 otherwise .
here we use @xmath128 .
the markov process is described by the ode : @xmath129 where @xmath130 .
the solution of is @xmath131 and its stationary state ( i.e. , @xmath132 ) is @xmath133 , where @xmath134 is the leading left eigenvector of @xmath135 .
a hard partition of the graph into @xmath136 communities can be encoded into the @xmath137 matrix @xmath138 , where @xmath139 if node @xmath7 belongs to community @xmath140 and zero otherwise .
the @xmath141 _ clustered autocovariance matrix _ of is @xmath142 and the entry @xmath143 of @xmath144 measures how likely it is that a random walker that started the process in community @xmath140 finds itself in community @xmath145 after time @xmath100 when at stationarity .
the diagonal elements of @xmath144 thus record how good the communities in @xmath138 are at retaining flows .
the _ markov stability of the partition _ is then defined as @xmath146 the optimised communities are obtained by maximising the cost function over the space of all partitions for every time @xmath100 to obtain an optimised partition @xmath147 .
this optimisation is np - hard ; hence with no guarantees of optimality . here we use the louvain greedy optimisation heuristic @xcite , which is known to give high quality solutions @xmath147 in an efficient manner .
the value of the markov time @xmath100 , i.e. the duration of the markov process , can be understood as a resolution parameter for the partition into communities @xcite . in the limit @xmath148 , markov stability
will assign each node to its own community ; as @xmath100 grows , we obtain larger communities because the random walkers have more time to explore the network @xcite .
we scan through a range of values of @xmath100 to explore the multiscale community structure of the network .
the code for markov stability can be found at github.com/michaelschaub/partitionstability . to identify the important partitions across time , we use two criteria of robustness @xcite .
firstly , we optimise 100 times for each value of @xmath100 and we assess the consistency of the solutions found .
a relevant partition should be a robust outcome of the optimisation , i.e. , the ensemble of optimised solutions should be similar as measured with the normalised variation of information @xcite : @xmath149 where @xmath150 is a shannon entropy and @xmath151 is the relative frequency of finding a node in community @xmath152 in partition @xmath153 .
we then compute the average variation of information of the ensemble of solutions from the @xmath154 louvain optimisations @xmath155 at each markov time @xmath100 : @xmath156 if all louvain runs return similar partitions , then @xmath157is small , indicating robustness of the partition to the optimisation . hence we select partitions with low values ( or dips ) of @xmath157 .
secondly , relevant partitions should also be optimal across markov time , as indicated by a low values of the cross - time variation of information : @xmath158 therefore , we also search for partitions with extended low value plateaux of @xmath159 @xcite .
m.b.d . acknowledges support from the james s. mcdonnell foundation postdoctoral program in complexity science / complex systems fellowship award ( # 220020349-cs / pd fellow ) , and the oxford - emirates data science lab .
g.b . acknowledges the support from the spanish ministry of economy fpi program ( bes-2012 - 053772 ) .
d.o . acknowledges support from an imperial college research fellowship and from the human frontier science program through a young investigator grant ( rgy0076 - 2015 ) .
j.p . acknowledges the support from the spanish ministry of economy and eu feder through the synbiofactory project ( cicyt dpi2014 - 55276-c5 - 1 ) .
m.b . acknowledges funding from the epsrc through grants ep / i017267/1 and ep / n014529/1 .
[ [ data - statement ] ] data statement : + + + + + + + + + + + + + + + no new data was generated during the course of this research .
a directed version of the rag could in principle be obtained from the boolean production / consumption matrices @xmath160 and @xmath161 as follows .
projecting onto the space of reactions gives the @xmath162 ( asymmetric ) adjacency matrix @xmath163 where the entries @xmath164 represent the total number of metabolites _ produced _ by reaction @xmath76 that are _ consumed _ by reaction @xmath50 .
a directed version of the reaction adjacency graph on @xmath3 nodes ( directly comparable to the standard rag ) is then @xmath165 clearly , when the metabolic model contains only reversible reactions , ( i.e. , the reversibility vector is all ones , @xmath166 ) , it follows that @xmath167 .
although @xmath168 does not include spurious edges introduced by non - existent backward reactions , its structure is still obscured by the effect of uninformative connections created by pool metabolites .
as an illustration of the graph construction , the toy metabolic network in fig . [
fig : toymodel_networks ] was taken from ref .
the graph matrices for this model are as follows : * reaction adjacency graph , eq . : @xmath169.\ ] ] * probabilistic flux graph , eq . :
@xmath170.\ ] ] * metabolic flux graph for fba scenario 1 , eq . : @xmath171.\ ] ] * metabolic flux graph for fba scenario 2 , , eq . : @xmath172.\ ] ]
a robust partition into seven communities in the rag was found at markov time @xmath173 ( fig . [ fig : markov_stability_template_networks]a ) . the communities at this resolution ( fig .
[ fig : avsd_comms]e ) are : * community c1@xmath86 contains all the reactions that consume or produce atp and water ( two pool metabolites ) .
production of atp comes mostly from oxidative phosphorylation ( atps4r ) and substrate level phosphorylation reactions such as phosphofructokinase ( pfk ) , phosphoglicerate kinase ( pgk ) and succinil - coa synthase ( sucoas ) .
reactions that consume atp include glutamine synthetase ( glns ) and atp maintenance equivalent reaction ( atpm ) .
the reactions l - glutamine transport via abc system ( glnabc ) , acetate transport in the form of phosphotransacetilase ( ptar ) , and acetate kinase ( ackr ) are also part of this community .
additionally , c1@xmath86 ( green ) contains also reactions that involve h@xmath174o . under normal conditions water
is assumed to be abundant in the cell , thus the biological link that groups these reactions together is tenuous .
* community c2@xmath86 includes the reactions nadh dehydrogenase ( nadh16 ) , cytochrome oxidase ( cytbd ) , and transport and exchange reactions .
these two reactions involve pool metabolites ( such as h@xmath87 ) which create a large number of connection .
other members include fumarate reductase ( fr7 ) and succinate dehydrogenase ( sucdi ) which couple the tca cycle with the electron transport chain ( through ubiquinone-8 reduction and ubiquinol-8 oxidation ) .
reactions that include export and transport of most secondary carbon sources ( such as pyruvate , ethanol , lactate , acetate , malate , fumarate , succinate or glutamate ) are included in the community as well .
these reactions are included in the community because of their influence in the proton balance of the cell .
most of these reactions do not occur under normal circumstances .
this community highlights the fact that in the absence of biological context , many reactions that do not normally interact can be grouped together .
* community c3@xmath86 contains reactions that produce or consume nicotinamide adenine dinucleotide ( nad@xmath87 ) , nicotinamide adenine dinucleotide phosphate ( nadp@xmath175 ) , or their reduced variants nadh and nadph .
the main two reactions of the community are nad(p ) transhydrogenase ( thd2 ) and nad@xmath87 transhydrogenase ( nadtrhd ) .
there are also reactions related to the production of nadh or nadph in the tca cycle such as isocitrate dehydrogenase ( icdhyr ) , 2-oxoglutarate dehydrogenase ( akgdh ) and malate dehydrogenase ( mdh ) .
the community also includes reactions that are not frequently active such as malic enzime nad ( me1 ) and malic enzime nadh ( me2 ) or acetate dehydrogenase ( acald ) and ethanol dehydrogenase ( alcd2x ) .
* community c4@xmath86 contains the main carbon intake of the cell ( glucose ) , the initial steps of glycolysis , and most of the pentose phosphate shunt .
these reactions are found in this community because the metabolites involved in these reactions ( e.g. , alpha - d - ribose-5-phosphate ( r5p ) or d - erythrose-4-phosphate ( e4p ) ) are only found in these reactions .
this community includes the biomass reaction due to the number of connections created by growth precursors .
* communities c5@xmath86 , c6@xmath86 and c7@xmath86 are small communities that contain oxygen intake , ammonium intake and acetaldehyde secretion reactions , respectively . and @xmath21 . * ( a ) communities in @xmath20 .
top plot : variation of information ( vi ) of the best partition found at markov time @xmath100 with every other partition at time @xmath176 .
bottom plot : number of communities and vi of the ensemble of solutions found at each markov time .
a robust partition into seven communities is found at @xmath173 .
( b ) communities and vi in @xmath21 .
a robust partition into five communities is found at @xmath177 . ] a robust partition into five communities in the pfg was found at markov time @xmath177 ( fig .
[ fig : markov_stability_template_networks]b ) .
the communities at this resolution ( fig .
[ fig : avsd_comms]c ) are : * community c1@xmath88 includes the first half of the glycolysis and the complete pentose phosphate pathway .
the metabolites that create the connections among these reactions such as d - fructose , d - glucose , or d - ribulose .
* community c2@xmath88 contains the main reaction that produces atp through substrate level ( pgk , pyk , ackr ) and oxidative phosphorylation ( atps4r ) .
the flow of metabolites among the reactions in this community includes some pool metabolites such as atp , adp , h@xmath1740 , and phosphate .
however , there are connections created by metabolites that only appear in a handful of reactions such as adenosine monophosphate ( amp ) whose sole producer is phosphoenolpyruvate synthase ( pps ) and its sole consumer is atps4r .
this community also contains the biomass reaction . *
community c3@xmath88 includes the core of the citric acid ( tca ) cycle such as citrate synthase ( cs ) , aconitase a / b ( aconta / b ) , and anaplerotic reactions such as malate synthase ( mals ) , malic enzyme nad ( me1 ) , and malic enzyme nadp ( me2 ) .
this community also includes the intake of cofactors such as co@xmath89 .
* community c4@xmath88 contains reactions that are secondary sources of carbon such as malate and succinate , as well as oxidative phosphorilation reactions .
* community c5@xmath88 contains some reactions part of the pyruvate metabolism subsystem such as d - lactate dehydrogenase ( ldh - d ) , pyruvate formate lyase ( pfl ) or acetaldehyde dehydrogenase ( acald ) .
in addition , it also includes the tranport reaction for the most common secondary carbon metabolites such as lactate , formate , acetaldehyde and ethanol .
this graph has 48 reactions with nonzero flux and 227 edges . at markov time
@xmath179 ( fig .
[ fig : markov_stability_fba_networks]a ) this graph has a partition into three communities ( fig .
[ fig : networksm]a ) : * community c1@xmath91 comprises the intake of glucose and most of the glycolysis and pentose phosphate pathway .
the function of the reactions in this community consists of carbon intake and processing glucose into phosphoenolpyruvate ( pep ) .
this community produces essential biocomponents for the cell such as alpha - d - ribose 5-phosphate ( rp5 ) , d - erythrose 4-phosphate ( e4p ) , d - fructose-6-phosphate ( f6p ) , glyceraldehyde-3-phosphate ( g3p ) or 3-phospho - d - glycerate ( 3pg ) .
other reactions produce energy atp and have reductive capabilities for catabolism .
* community c2@xmath91 contains the electron transport chain which produces the majority of the energy of the cell . in the core _
e coli _ metabolic model the chain is represented by the reactions nadh dehydrogenase ( nadh16 ) , cytochrome oxidase bd ( cytbd ) and atp synthase ( atps4r ) .
this community also contains associated reactions to the electron transport such as phosphate intake ( expi(e ) , pit2 ) , oxygen intake ( exo2(e ) , o2 t ) and proton balance ( exh(e ) ) .
this community also includes the two reactions that represent energy maintenance costs ( atpm ) , and growth ( biomass ) ; this is consistent with the biological scenario because atp is the main substrate for both atpm , and the biomass reaction . *
community c3@xmath91 contains the tca cycle at its core .
the reactions in this community convert pep into atp , nadh and nadph .
in contrast with c1@xmath91 , there is no precursor formation here . beyond the tca cycle , pyruvate kinase ( pyk ) , phosphoenolpyruvate carboxylase
( ppc ) and pyruvate dehydrogenase ( pdh ) appear in this community .
these reactions highlight the two main carbon intake routes in the cycle : oxalacetate from pep through phosphoenol pyruvate carboxylase ( ppc ) , and citrate from acetyl coenzyme a ( acetyl - coa ) via citrate synthase ( cs ) .
furthermore , both routes begin with pep , so it is natural for them to belong to the same community along with the rest of the tca cycle . likewise , the production of l - glutamate from 2-oxoglutarate ( akg ) by glutamate dehydrogenase ( gludy ) is strongly coupled to the tca cycle .
* community c1@xmath93 in this graph is similar to its counterpart in @xmath90 , but with important differences .
for example , the reactions in charge of the glucose intake ( exglc(e ) and glcpts ) are no longer part of the network ( i.e. , they have zero flux ) , and reactions such as malic enzyme napd ( me2 ) and phosphoenolpyruvate caboxykinase ( ppck ) , which now appear in the network , belong to this community .
this change in the network reflect the cell s response to a new biological situation .
the carbon intake through ethanol has changed the direction of glycolysis into gluconeogenesis @xcite ( the reactions in c1@xmath91 in fig . [
fig : networksm]a are now operating in the reverse direction in fig . [
fig : networksm]b ) .
the main role of the reactions in this community is the production of bioprecursors such as pep , pyruvate , 3-phospho - d - glycerate ( 3pg ) glyceraldehyde-3-phosphate ( g3p ) , d - fructose-6-phosphate ( f6p ) , and d - glucose-6-phosphate , all of which are substrates for growth .
reactions me2 and ppck also belong to this community due to their production of pyr and pep .
reactions that were in a different community in @xmath90 , such as gludy and icdhyr which produce precursors l - glutamate and nadph respectively , are now part of c1@xmath93 .
this community also includes the reactions that produce inorganic substrates of growth such as nh@xmath113 , co@xmath174 and h@xmath89o . *
community c2@xmath93 contains the electron transport chain and the bulk of atp production , which is similar to c2@xmath91 .
however , there are subtle differences that reflect changes in this new scenario .
ethanol intake and transport reactions ( exetoh(e ) and etoht2r ) appear in this community due to their influence in the proton balance of the cell .
in addition , c2@xmath93 contains nadp transhydrogenase ( thd2 ) which is in charge of nadh / nadph balance .
this reaction is present here due to the nad consumption involved in the reactions acald and ethanol dehydrogenase ( alcd2x ) , which belong to this community as well .
* community c3@xmath93 contains most of the tca cycle .
the main difference between this community and c1@xmath91 is that here acetyl - coa is extracted from acetaldehyde ( which comes from ethanol ) by the reaction acetaldehyde dehydrogenase reaction ( acald ) , instead of the classical pyruvate from glycolysis .
the glycoxylate cycle reactions isocitrate lyase ( icl ) and malate synthase ( mals ) which now appear in the network , also belong to this community .
these reactions are tightly linked to the tca cycle and appear when the carbon intake is acetate or ethanol to prevent the loss of carbon as co@xmath174 .
* community c1@xmath95 contains the reactions responsible d - glucose intake ( exglc ) and most of the glycolysis . the reaction that represents the cellular maintenance energy cost , atp maintenance requirement ( atpm ) ,
is included in this community because of the increased strength of its connection to the substrate - level phosphorilation reaction phosphoglycerate kinase ( pgk ) .
also note that reactions in the pentose phosphate pathway do not belong to the same community as the glycolysis reactions ( unlike in @xmath90 and @xmath92 ) .
* community c2@xmath95 contains the conversion of pep into formate through the sequence of reactions pyk , pfl , forti and exfor(e ) .
more than half of the carbon secreted by the cell becomes formate .
* community c3@xmath95 includes the biomass reaction and the reactions in charge of supplying it with substrates .
these reactions include the pentose phosphate pathway ( now detached from c1@xmath91 ) , which produce essential growth precursors such as alpha - d - ribose-5-phosphate ( r5p ) or d - erythrose-4-phosphate ( e4p ) . the tca cycle is present as well because its production of two growth precursors : 2-oxalacetate and nadph . finally , the reactions in charge of acetate production ( ackr , act2r and exac(e ) ) are also members of this community through the ability of ackr to produce atp .
glutamate metabolism reaction gludy is also included in this community .
it is worth mentioning that the reverse of atp synthase ( atps4r ) is present in this community because here , unlike in @xmath90 , atps4r consumes atp instead of producing it . when this flux is reversed , then atps4r is in part responsible for ph homeostasis .
* community c4@xmath95 includes the main reactions involved in nadh production and consumption , which occurs via glyceraldehyde-3-phosphate dehydrogenase ( gapd ) .
nadh consumption occurs in two consecutive steps in ethanol production : in acald and alcd2x .
the phosphate intake and transport reactions expi(e ) and pit2r belong to this community because most of the phosphate consumption takes place at gapd .
interestingly , the core reaction around which the community forms ( gapd ) is not present in the community .
it is included in earlier markov times but when communities start to get larger the role of gapd becomes more relevant as a part of the glycolysis than its role as a nadh hub .
this is a good example of how the graph structure and the clustering method are able to capture two different roles in the same metabolite .
has a robust partition into three communities at @xmath179 .
( b ) @xmath180 has a partition into three communities at @xmath177 .
( c ) @xmath181 has four communities at @xmath173 .
( d ) @xmath182 has three communities at @xmath183 . ] * community c1@xmath97 contains the glycolysis pathway ( detached from the pentose phosphate pathway ) .
this community is involved in precursor formation , atp production , substrate - level phosphorylation and processing of d - glucose into pep .
* community c2@xmath97 contains the bioenergetic machinery of the cell ; the main difference to the previous scenarios is that the electron transport chain has a smaller role in atp production ( atps4r ) , and substrate - level phosphorylation ( pgk , pyk , sucoas , ackr ) becomes more important . in @xmath96
the electron transport chain is responsible for the 21.8% of the total atp produced in the cell while in @xmath90 it produces 66.5% .
the reactions in charge of intake and transport of inorganic ions such as phosphate ( expi(e ) and pit2r ) , o@xmath174 ( exo@xmath174(e ) and o@xmath174t)and h@xmath89o ( exh@xmath174o and h@xmath174ot ) belong to this community as well .
this community includes the reactions in the pentose phosphate pathway that produce precursors for growth : transketolase ( tkt2 ) produces e4p , and ribose-5-phosphate isomerase ( rpi ) produces r5p .
* community c3@xmath97 is the community that differs the most from those in the other aerobic growth networks ( @xmath90 and @xmath92 ) .
this community gathers reactions that under normal circumstances would not be so strongly related but that the limited availability of ammonium and phosphate have forced together ; its members include reactions from the tca cycle , the pentose phosphate pathway , nitrogen metabolism and by - product secretion .
the core feature of the community is carbon secretion as formate and acetate .
reactions ppc , malate dehydrogenase ( mdh ) reverse and me2 channel most of the carbon to the secretion routes in the form of formate and acetate .
the production of l - glutamine seems to be attached to this subsystem through the production of nadph in me2 and its consumption in the glutamate dehydrogenase napd ( gludy ) . ma hw , zhao xm , yuan yj , zeng ap ( 2004 ) decomposition of metabolic network into functional modules based on the global connectivity structure of reaction graph . bioinformatics ( oxford , england ) 20 : 18706 .
sauer u , lasko dr , fiaux j , hochuli m , glaser r , et al .
( 1999 ) metabolic flux ratio analysis of genetic and environmental modulations of escherichia coli central carbon metabolism .
journal of bacteriology 181 : 6679 - 6688 . king za , drger a , ebrahim a , sonnenschein n , lewis ne , et al . ( 2015 ) escher : a web application for building , sharing , and embedding data - rich visualizations of biological pathways .
plos comput biol 11 : 1 - 13 .
folch - fortuny a , tortajada m , prats - montalbn jm , llaneras f , pic j , et al .
( 2015 ) mcr als on metabolic networks : obtaining more meaningful pathways . chemometrics and intelligent laboratory systems 142 : 293303 .
schilling ch , letscher d , palsson bo ( 2000 ) theory for the systemic definition of metabolic pathways and their use in interpreting metabolic function from a pathway - oriented perspective .
j theor biol 203 : 229248 .
page l , brin s , motwani r , winograd t ( 1999 ) the pagerank citation ranking : bringing order to the web .
technical report 1999 - 66 , stanford infolab .
http://ilpubs.stanford.edu:8090/422/. previous number = sidl - wp-1999 - 0120 .
delvenne jc , schaub mt , yaliraki sn , barahona m ( 2013 ) the stability of a graph partition : a dynamics - based framework for community detection . in :
mukherjee a , choudhury m , peruani f , ganguly n , mitra b , editors , dynamics on and of complex networks , volume 2 , springer new york , modeling and simulation in science , engineering and technology .
221 - 242 .
colijn c , brandes a , zucker j , lun ds , weiner b , et al .
( 2009 ) interpreting expression data with metabolic flux models : predicting _ mycobacterium tuberculosis _
mycolic acid production .
plos comput biol 5 : e1000489 . | cells adapt their metabolic fluxes in response to changes in the environment .
we present a systematic flux - based framework for the construction of graphs to represent organism - wide metabolic networks .
our graphs encode the directionality of metabolic fluxes via links that represent the flow of metabolites from source to target reactions .
the methodology can be applied in the absence of a specific biological context by modelling fluxes as probabilities , or tailored to different environmental conditions by incorporating flux distributions computed from constraint - based modelling such as flux balance analysis .
we illustrate our approach on the central carbon metabolism of _ escherichia coli _ and study the derived graphs under various growth conditions .
the results reveal drastic changes in the topological and community structure of the metabolic graphs , which capture the re - routing of metabolic fluxes under each growth condition . by integrating constraint - based models and tools from network science ,
our framework allows for the interrogation of environment - specific metabolic responses beyond fixed , standard pathway descriptions . |
dendrimers , characterized by regular branching with radial symmetry , represent a unique class of polymers possessing a large number of functional terminal groups in their outermost dendritic shell.@xcite these `` dream molecules''@xcite have recently attracted a lot of attention due to potential applications in technology and medicine , see , for example , refs . and references therein .
polyamidoamine ( pamam ) dendrimers , in particular , are commonly used molecules in this field .
they are protonated under physiological conditions,@xcite thus their high surface charge coupled to their ability to adapt conformation in response to changes in their surrounding environment make pamam dendrimers very attractive candidates for biomedical applications such as drug and gene delivery.@xcite thus far , most theoretical and computational investigations have focused on neutral dendrimers.@xcite since the pioneering paper of welch and muthukumar,@xcite there have been only a few computational studies of charged model dendrimers.@xcite while these studies have provided a great deal of insight into the properties of dendrimers , it is obvious that their scope has been limited due to a major computational cost associated with modeling of complex dendrimers in a hydrodynamic solvent .
hence , the previous studies have treated the groups of a charged dendrimer using the linearized poisson
boltzmann ( debye hckel ) theory , and the solvent as a continuous medium .
despite these limitations , the above simulations demonstrated that the strength of electrostatic interactions may have a strong effect on the dendrimer s size as well as on other structural and dynamic properties .
in particular , it was shown that the size of any charged dendrimer increases for an increasing effective screening length which corresponds to a decreasing salt concentration.@xcite as for more accurate treatments , lee et al.@xcite simulated pamam dendrimers at an atomic resolution .
they found that a pamam dendrimer swells significantly when ph decreases from a high ph value ( no amines are protonated ) to neutral ( all the primary amines are protonated ) , and to low ph ( all the primary amines and tertiary amines are protonated ) .
however , they treated the counterions as well as the solvent implicitly ( except for dendrimers of second generation which were solvated in explicit water).@xcite a more realistic atomic - scale simulation of pamam dendrimers was performed very recently by goddard et al.,@xcite who accounted for water molecules and counterions in an explicit manner .
they showed that the presence of solvent leads to a swelling of the dendrimer , while the effect of ph was found to be similar to that reported in earlier studies.@xcite as a part of counterions were found to condense onto the dendrimer,@xcite it was stated that the presence of counterions increases the swelling of a dendrimer .
the effect of counterions on dendrimer charges and related structural aspects were not clarified , though , since for the counterion concentrations they used , the amine charges were strongly screened : the corresponding debye length was about 34 and hence close to the minimal value considered by welch and muthukumar.@xcite this implies that the effect of smaller counterion concentrations and , hence , the overall role of counterions has remained unresolved . in general
, there is reason to emphasize that the applicability of the debye
hckel approximation for dendrimers should not be taken for granted .
it is valid in the dilute limit here it means the situation where the separation between charges exceeds the bjerrum length .
however , that condition breaks down easily in dense structures such as dendrimers , which implies that in most of the cases counterions must be taken _
explicitly _ into account .
computational studies of flexible linear polyelectrolytes have demonstrated@xcite that the presence of explicit counterions is indeed crucial : increasing the strength of electrostatic interactions can provoke counterion condensation and , therefore , lead to dramatic changes in the polymer s structure and dynamics .
it is reasonable to expect similar counterion effects also for dendrimers
. a recent computational study of model dendrimers suggests that this is indeed the case.@xcite in this study we aim to clarify the role of counterions .
we have performed extensive molecular dynamics simulations using a generic cationic model dendrimer of the fourth generation .
counterions and solvent molecules are _ explicitly _ included .
a systematic variation of the strength of electrostatic interactions have allowed us to examine the influence of counterions on the structural and dynamical properties of charged dendrimers .
it turns out that the size of charged dendrimers depends _ non - monotonically _ on the strength of electrostatic interactions in the system clearly indicating a pronounced effect of counterion condensation .
a sketch of the dendritic topology used in this study .
beads of the last , or the terminal , generation shell ( the illustration is for a dendrimer of the second generation ) are positively charged ( black ) . in the simulations we used dendrimers of the fourth generation.,width=226 ] we performed md simulations of dendrimers of the 4th generation using the freely jointed `` bead - and - spring '' model .
each bead of the last , or the terminal , generation was set to carry a positive unit charge .
the dendrimer was solvated in a box of non - polar molecules ( beads ) and an appropriate number of counterions was added to preserve overall charge neutrality .
the bonds were modeled as harmonic springs with their potential energy given by @xmath0 , where @xmath1 is the reference bond length .
the spring constant @xmath2 was chosen to be rather high as has been done in earlier studies.@xcite all short - range non - electrostatic non - bonded interactions were described by the lennard
jones potential @xmath3 $ ] with @xmath4 .
bead interaction diameter @xmath5 was taken to be the same for all beads in the system .
the lennard
jones potential was cut off at @xmath6 and shifted to zero.@xcite the dendrimer was in a good solvent .
the long - range electrostatic interactions between the charged terminal beads ( with a positive unit charge on each ) and negatively charged counterions were handled using the particle - mesh ewald ( pme ) method,@xcite which has been shown to perform well in soft matter simulations.@xcite the strength of electrostatic interactions is characterized by the so - called bjerrum length @xmath7 where @xmath8 is the dielectric permittivity of the solvent . to examine the effects of counterions on the dendrimer s structure and dynamics ,
the bjerrum length @xmath9 was systematically varied from 0 to @xmath10 .
this change in @xmath9 may be interpreted as a variation of the dielectric properties of the solvent ( @xmath11 ) . in all
, we studied 10 different systems with @xmath12 taken to be 0.0 ; 0.2 ; 0.4 ; 0.6 ; 0.8 ; 1.0 ; 2.0 ; 3.2 ; 6.4 ; and 8.0 .
the dendritic topology employed in our study is schematically depicted in fig .
we studied a trifunctional dendrimer of the fourth generation with single bonds between the branch points .
the dendrimer had 94 beads of which 48 were terminal ones . to generate initial configurations ,
the procedure of murat and grest was used.@xcite the dendrimer was then solvated in a cubic box with 4692 solvent particles .
that corresponds to a polymer fraction of about 0.02 .
the density was set to 1.688@xmath13 ( or , alternatively , to 0.864@xmath14 ) , similar to refs . and .
the linear size of the simulation box size was set to 14.16@xmath15 , and periodic boundary conditions were applied in all three dimensions .
the ratio of the average dendrimer size to the simulation box size was about 0.2 . for simulating charged dendrimers ,
unit positive charges were assigned to all beads of the terminal g4 generation shell of the neutral dendrimer .
this corresponds to physiological ( neutral ) ph conditions when the primary amines of a pamam dendrimer become protonated.@xcite to keep the system electroneutral , randomly chosen 48 solvent beads were converted to counterions by assigning negative unit charges on them .
this system was used as the starting configuration for md simulations of charged dendrimers .
all systems with different bjerrum lengths were first equilibrated for @xmath16 time steps in the nvt ensemble with the time step @xmath17 , where @xmath18 is the characteristic time of the model ( @xmath19 is the mass of a bead ) .
temperature was controlled using the berendsen scheme,@xcite while all the bond lengths were kept constant by using the shake algorithm.@xcite after equilibration , data were collected from production runs of @xmath20 time steps each in the nve ensemble . the time step in the production runs
was set to @xmath21 ( constraints were not applied to dendrimer s bonds during the production ) .
all the simulations were performed using the gromacs package.@xcite each simulation was run on a single 3.2 ghz pentium-4 processor .
average mean - square radius of gyration @xmath22 of the whole dendrimer ( solid circles ) , and of the beads of the _ terminal _ dendrimer shell ( open circles ) as a function of the bjerrum length @xmath9 .
error bars are of the same size as the symbols .
the error bars were estimated as the standard errors of mean by splitting trajectories into 10 pieces of 10@xmath23 time steps each.,width=302 ] ( color online ) .
typical configurations of charged dendrimers at @xmath24 ( top ) and @xmath25 ( bottom ) .
counterions are shown as spheres .
solvent is omitted for clarity .
, title="fig:",width=226 ] ( color online )
. typical configurations of charged dendrimers at @xmath24 ( top ) and @xmath25 ( bottom ) .
counterions are shown as spheres .
solvent is omitted for clarity .
, title="fig:",width=226 ] _ top _ : average mean - square radius of gyration @xmath22 calculated for counterions with respect to the center of mass ( cm ) of the dendrimer .
_ bottom _ : average fraction of counterions condensed onto the charged dendrimer as a function of the bjerrum length @xmath9 .
see text for the criterion used to identify condensed ions.,width=302 ] average coordination numbers of ions with the dendrimer s beads ( solid circles ) , with the charged beads of the g4 shell ( open circles ) , and with solvent beads ( open squares ) as a function of the bjerrum length @xmath26.,width=302 ] to examine possible finite size effects , we performed additional simulation runs using simulations containing twice the number of solvent molecules .
we examined the box size effects for a neutral dendrimer as well as for the charged dendrimers with @xmath27 and @xmath25 . for the case of neutral dendrimers we did not find any noticeable deviations when the larger box size was employed .
for charged dendrimers , the average size of the counterion cloud around the dendrimer was found to naturally increase upon increasing the box size .
this , however , did not produce any considerable changes in the structural or dynamic properties of dendrimers .
as in previous computational studies,@xcite initial equilibration was monitored through the time evolution of the mean - square radius of gyration @xmath28 where @xmath29 is the number of beads and @xmath30 is the position of the center of mass ( cm ) of the dendrimer .
this quantity , being a measure of the overall size of the dendrimer , was found to remain stable during the entire production run .
the above , however , is not enough for a system containing unscreened charges . is it further required that the ions have reached equilibrium . to monitor the equilibration of ions , we first calculated the radial distribution functions ( rdfs ) for ion ion , ion dendrimer bead , and ion solvent molecule pairs ( data not shown ) .
the position of the first minimum defines the radius of the first coordination shell of a counterion .
this procedure has been used in other related studies as well.@xcite the radius was found to be @xmath311.3@xmath15 .
the coordination numbers were then calculated by counting the total numbers of beads within the first coordination shell .
the coordination numbers are found to become stable as a function of time during the equilibration period ( see sect .
[ model ] ) .
next , we describe the structural characteristics of charged dendrimers at different strengths of electrostatic interactions . in fig .
[ rg ] we plot the average mean - square radius of gyration @xmath22 of the dendrimer as a function of bjerrum length @xmath9 .
remarkably , the dependence of @xmath22 on @xmath9 is found to be _ non - monotonic_. increasing the bjerrum length from zero leads to a pronounced swelling of the dendrimer , the maximum being at @xmath32 as shown in fig .
further increase is found to have an opposite effect , i.e. , the dendrimer shrinks .
such a behavior resembles the findings for linear@xcite and dendritic@xcite polyelectrolytes , and is associated with counterion condensation .
figure [ snap ] shows snapshots from our simulations for the swollen and shrunk dendrimers .
the @xmath9-dependence for the radius of gyration calculated for charged beads of the terminal shell ( @xmath33 ) follows closely that of the whole dendrimer ( @xmath34 ) , see fig .
interestingly , the @xmath33 exceeds the radius of gyration of the dendrimer at all bjerrum lengths . for neutral dendrimers
this finding is in agreement with experimental data@xcite and with brownian dynamics simulations.@xcite as demonstrated in ref . , and
as we proceed to show , the fact that @xmath35 does _ not _ mean that beads of the terminal generation shell are localized near the dendrimer s periphery . to clarify the role of counterions , we calculated the mean - square radius of gyration @xmath36 of the counterions , see fig .
[ ions ] ( top ) . that serves as a measure for the size of the counterion cloud around the dendrimer .
it turns out that the ion cloud becomes smaller as the bjerrum length increases .
this decrease , however , is found to be monotonic .
furthermore , in fig . [ ions ] ( bottom ) we estimate the fraction of counterions condensed onto the dendrimer . for doing that we used a simple criterion , namely we counted counterions which had dendrimer beads in their first coordination shells . as fig .
[ ions ] shows , the number of condensed ions increases continuously with the strength of electrostatic interactions in the system .
the above suggests that there is a subtle interplay between two opposite effects upon increasing @xmath9 : ( _ i _ ) a growth in the strength of the _ repulsive _ electrostatic interactions between the charged terminal beads and ( _ ii _ ) a strengthening of condensation of counterions due to the _ attractive _ interactions between counterions and the oppositely charged dendrimer beads .
the former leads to a swelling of the charged dendrimer and the latter screens positive charges of the terminal dendrimer generation giving rise to shrinking of the dendrimer .
the maximum of the dendrimer s size is found to be at @xmath32 ( i.e. at @xmath37 ) , see fig .
this corresponds to swelling by @xmath3112.5% with respect to the neutral case .
condensation of counterions onto the dendrimer should be accompanied by a decrease in its hydration , which is characterized by the number of solvent molecules in the ion s first coordination shell .
indeed , that can be seen by computing the average coordination numbers of counterions with dendrimer s beads and with solvent molecules , see fig .
the increase of the strength of electrostatic interactions in the system provokes a condensation of counterions and , correspondingly , leads to a loss of solvent molecules from the ions first coordination ( hydration ) shell ( fig .
[ cn ] ) . to analyze the locations of the dendrimer s beads , solvent molecules , and counterions in more detail , we calculated componentwise number density profiles @xmath38 as a function of the radial distance @xmath39 from the center of mass ( cm ) of the dendrimer , see fig .
[ density ] ( fig .
[ beadnum ] shows the corresponding radial bead distributions ) .
for a neutral dendrimer the density profiles are in agreement with earlier molecular dynamics studies of dendrimers in an explicit solvent.@xcite the overall number density of a dendrimer develops a plateau at @xmath40 from its cm . the beads of the terminal shell are broadly distributed , implying a considerable degree of back - folding .
the solvent molecules penetrate deeply into the dendrimer and their density profile has a plateau at @xmath40 , see fig .
[ density ] . at small bjerrum lengths ( @xmath41 )
the charged dendrimer swells and the density profile plateau broadens considerably with a simultaneous drop in the plateau s magnitude , see , e.g. , fig . [ density ] for @xmath24 .
the charged beads of the terminal shell are re - distributed towards the dendrimer s periphery making the interior of the dendrimer significantly more accessible for solvent and counterions .
further increase in bjerrum length leads to a shrinkage since counterion condensation starts to dominate .
therefore , the dendrimer s density profile changes towards a similar profile as observed at small @xmath9 with one important exception : the amount of condensed ions increases considerably , and their density profiles overlap more and more with the density profiles of the charged terminal beads of the dendrimer ( up to almost complete coincidence of the profiles in the case of extremely large bjerrum lengths , see fig . [ density ] for @xmath42 ) . this shrinkage is accompanied by a reduction of voids in the dendrimer s interior leading to squeezing of solvent molecules out of the dendrimer , see figs . [ density ] and [ beadnum ] .
next , we take a look at the screening of the dendrimer s charges by counterions . in fig .
[ charge ] we plot the charge density profiles @xmath43 for the whole system as a function of the radial distance @xmath39 from the cm of the dendrimer .
we recall that since our solvent is non - polar , the charge densities @xmath43 are defined only by the interplay between the positively charged terminal beads and the negatively charged counterions .
figure [ charge ] clearly demonstrates screening : the height of the positive @xmath43-peaks drops drastically with increasing bjerrum length as compared to the @xmath43-peak for @xmath44 when the counterion condensation has just started .
while the main peaks of all @xmath43-curves are located in the domain of the positive charges , starting with @xmath45 , one observes small minima close to the cm of the dendrimer where @xmath43 becomes negative , see fig .
[ charge ] .
we address these minima to the fact that some of condensed counterions can be localized deep in the dendrimer s interior rather than beside charged terminal beads giving rise to an excess of negative charges close to the dendrimer s center of mass , see figs .
[ density ] and [ beadnum ] . in turn , on the right of @xmath43-peaks we see another region where the overall charge of the system is also negative .
this region of negative charge , being associated with `` unbound '' counterions in bulk solvent , becomes smaller with @xmath9 and eventually disappears at large bjerrum lengths .
radial charge densities @xmath43 as a function of the radial distance @xmath39 from the cm of a dendrimer for @xmath26 equal to 0.2@xmath15 ( solid line ) , 0.4@xmath15 ( dashed line ) , 0.8@xmath15 ( solid line with circles ) , and 3.2@xmath15 ( solid line with stars ) . to reduce the noise in the data , the charge densities shown here were first fitted to splines.@xcite , width=302 ]
the molecular dynamics simulation technique provides an access to the dynamic properties , and next we discuss the dynamics of charged dendrimers in a salt - free solution .
first , we consider the relaxation of vectors pointing from the center of mass ( cm ) of the dendrimer to the beads of a given generation shell.@xcite if the terminal ( g4 ) generation shell is considered , the corresponding autocorrelation function ( acf ) of `` cm - g4 '' vector can be employed to analyze the global reorientational behavior of the whole dendrimer.@xcite it is , however , necessary to bear in mind that there is no reason for the beads of the last generation to be located at the dendrimer s periphery . in general , relaxation of the `` cm - g4 '' vector is determined by the rotation of the whole dendrimer together with fluctuations in the vector s length . to characterize the dendrimer s rotation
, we calculated the acf of the _ unit _ vector directed along the `` cm - g4 '' vector .
it turns out that the corresponding relaxation times @xmath46 are very close to the relaxation time of the `` cm - g4 '' vector ( @xmath47 ) .
thus , the relaxation of the `` cm - g4 '' vector is governed almost exclusively by the rotation of the dendrimer in agreement with the results from brownian dynamics simulations of dendrimers with rigid bonds.@xcite characteristic relaxation times @xmath47 and @xmath46 related to the dendrimer s reorientational dynamics as extracted from the corresponding autocorrelation functions of vectors pointing from the cm of the dendrimer to the beads of its terminal g4 generation shell , see text for details .
error bars are of the order of the symbol size .
, width=302 ] figure [ taureo ] shows that the global reorientational motion of the charged dendrimer depends non - monotonically on the bjerrum length .
such a behavior resembles the @xmath9-dependence of @xmath22 as was seen earlier in fig .
this is expected since it is reasonable to assume that the global reorientational dynamics is controlled by the dendrimer s size : larger size requires more time for reorientation . as fig .
[ taureo ] shows , there is a sharp maximum at @xmath32 .
it is also interesting that although at large bjerrum lengths the average size of the charged dendrimer is found to be smaller than the size of a neutral dendrimer ( fig .
[ rg ] ) , that is not reflected in the characteristic relaxation times shown in fig .
[ taureo ] .
the former can be explained by the inhomogeneous distribution of counterions inside the dendrimer , giving rise to the additional electrostatic attraction between different parts of the dendrimer interior . on the other hand ,
the global reorientation dynamics is not only governed by the dendrimer s size but also by its mass . at @xmath48
almost all counterions are bound to the dendrimer ( fig .
[ ions ] ) , thus increasing the dendrimer s effective mass by a factor of 1.5 in comparison to the neutral one .
this , in turn , slows down the global reorientational motion of a dendrimer .
time evolution of the normalized autocorrelation function @xmath49 of bond vectors @xmath50 of the innermost ( _ top _ ) and the outermost ( _ bottom _ ) shells .
the legend shows the different bjerrum lengths used .
, width=302 ] as seen in fig .
[ taureo ] , the global reorientational dynamics is determined by rather long characteristic times up to @xmath31200@xmath51 .
the local orientational motions are characterized by much shorter times . as an example
, we consider here the re - orientation dynamics of the bond vectors @xmath50 . in fig .
[ p1 ] we plot the time evolution of the corresponding autocorrelation function @xmath49 for the innermost ( g0 ) and the outermost ( g4 ) shells .
we note that the relaxation of @xmath52 can not be characterized by a single relaxation time .
therefore , we focus here on the time behavior of @xmath52 rather than on characteristic times .
the reorientation of a bond slows down with the size of the dendrimer , see fig .
[ p1 ] ( top ) .
such a behavior is found to hold for all inner generation shells from g1 to g3 ( data not shown ) . remarkably , the above non - monotonic ( notice the order of the curves ) dependence on @xmath9 in fig .
[ p1 ] ( top ) breaks down for @xmath52 of the terminal ( outermost ) generation shell as it is seen in fig .
[ p1 ] ( bottom ) .
this feature can be linked to condensation of counterions which plays a dominant role at @xmath53 and , therefore , can hinder the reorientation of bonds of the terminal g4 shell .
the characteristic relaxation time @xmath54 for a vector pointing from the cm of a dendrimer to its core as a function of the bjerrum length @xmath9 .
error bars are of the order of symbol size.,width=302 ] the local re - orientation dynamics of dendrimer bonds can also be characterized by the autocorrelation function @xmath55 .
it turns out that the simple relation @xmath56 holds for all bonds in agreement with the previous simulation studies.@xcite last , we focus on the relaxation of the vector pointing from the dendrimer s center of mass to its core ( `` cm - core '' vector ) .
@xcite the core and the cm are typically found to be very close to each other .
however , they do not exactly coincide and the relaxation of the corresponding `` cm - core '' vector is characterized by small ( but finite ) relaxation times . in fig .
[ cm ] we plot the corresponding characteristic relaxation times @xmath54 . as argued above , the relaxation times turn out to be very small , just a few @xmath57 .
the @xmath9-dependence of the @xmath54 resembles the behavior of the @xmath22-curve in fig .
[ rg ] , i.e. , increase in the dendrimer size speeds up the relaxation of the `` cm - core '' vector and _ vice versa_. this effect can be explained in terms of the mobility of a dendrimer core : larger dendrimers are characterized by smaller densities of their interior ( see fig . [ density ] ) and , therefore , by more mobile cores which relax faster .
we have performed a systematic molecular dynamics study of generic cationic dendrimers solvated in _
explicit _ solvent in the presence of _ explicit _ counterions .
the main goal was to understand to what degree the structural and dynamic properties of charged dendrimers are sensitive to the strength of electrostatic interactions , and what is the role of counterions .
we found an intriguing interplay between the _ repulsive _ interactions of the dendrimer s charged terminal beads and their _ attractive _ interactions with oppositely charged counterions .
depending on the type of dominant interactions , one observes swelling or shrinking of the molecule as the strength of the electrostatic interactions is increased . in other words ,
the size of a charged dendrimer depends _ non - monotonically _ on the bjerrum length @xmath9 as shown in fig .
most of the dendrimer s structural and dynamic characteristics considered in this study were found to follow this pattern . for small bjerrum lengths a charged dendrimer swells with @xmath9 and
its size reaches maximum at @xmath58 . at larger bjerrum lengths the condensation of counterions dominates and the dendrimer shrinks .
the counterion condensation is often discussed in the framework of the manning theory,@xcite which predicts the condensation to occur when @xmath59 where @xmath60 is the distance between charges along the polymer backbone .
however , the manning s criterion is not applicable in our case .
the manning s theory is formulated for linear polyelectrolytes and not for spherical - like branched objects such as dendrimers .
second , the charged dendrimer considered in our study has only its terminal beads charged , making the meaning of the parameter @xmath60 uncertain .
the counterion condensation , the effects of valence , charging , and the size and shape of the dendrimer could more appropriately be studied using the method of patra et al.@xcite who studied the charging of spherical three dimensional objects under different conditions .
a detailed study of the above factors is , however , beyond the current work and will be addressed in a future publication .
the bjerrum length @xmath61 corresponds to the position of the @xmath22-peak , see fig .
[ rg ] , and , therefore , separates swelling and shrinking regimes . remarkably , on the left - hand side of the peak , the separation between charges _ always _ exceeds the bjerrum length because the beads can not get closer than @xmath62 and @xmath63 . on the other hand ,
the counterion condensation is rather weak in this @xmath9-domain .
therefore , one can expect the debye hckel approximation to be valid at around @xmath64 and to break down at larger bjerrum lengths .
it is instructive to relate the dendrimer bond length @xmath15 ( and , correspondingly , the average dendrimer size @xmath65 and the bjerrum length @xmath9 ) to the characteristic length scales typical for real dendrimers in aqueous solutions .
the value of 17.1 is available from small angle x - ray scattering experiments@xcite for the average radius of gyration @xmath65 of a pamam dendrimer of the fourth generation .
a very recent md study@xcite reported @xmath66 for a non - protonated pamam dendrimer in explicit water .
in another study , lee et al .
@xcite reported @xmath67 .
that was obtained from md studies of pamam dendrimers without solvent ( @xmath68 in ref .
was obtained under the same conditions ) .
assuming that dendrimer swells on about 15% in water,@xcite the results of lee et al.@xcite seem to be in accord with other studies .
we took @xmath69 for our parameterization . for a neutral dendrimer we have @xmath70 ,
see fig .
[ rg ] , such that @xmath71 .
thus , the critical bjerrum length corresponding to the dendrimer s maximal size can be estimated to be 5.4 . in turn , the bjerrum length in water at 300 k is about 7.1 ( or @xmath72 ) , i.e. , it exceeds the critical value . in practice , since the @xmath22-peak is not very sharp , the value of @xmath22 for a dendrimer in _ water _
, being located next to the peak , is rather close to the maximum of @xmath22 , see fig .
[ rg ] . taking also into account the uncertainty in defining the value of @xmath15
, one can conclude that counterion condensation will not affect significantly pamam dendrimers in water under physiological conditions .
to summarize , we demonstrated that including explicit counterions can have a dramatic effect on the structure and dynamics of charged dendrimers , and under certain conditions they can not be treated implicitly .
therefore , simplified approaches , such as the well - known debye
hckel approximation , have to be applied with great care . given that the most exciting applications of protonated dendrimers are related to drug and gene delivery , it would be interesting to clarify the role of counterions in complexation of charged dendrimers with linear polyelectrolytes bearing opposite charges .
@xcite this problem is addressed in our ongoing studies .
fruitful discussions with dr . alexey v. lyulin are gratefully acknowledged .
this work has been supported by the academy of finland through its center of excellence program ( a.a.g .
, i.v . ) and through grants 202598 ( a.a.g . ) , 80246 ( i.v . ) , 00119 ( m.k . ) , 54113 ( m.k . ) , and 211579 ( s.v.l . ) and by the russian foundation of basic research through grants 05 - 03 - 32332 ( a.a.g . ) and 05 - 03 - 32450-a ( s.v.l . ) .
s.v.l . acknowledges the support of the government of st.petersburg , russia ( grant pd 05 - 1.3 - 101 ) and m.k .
the support from emil aaltonen foundation ( finland ) .
the simulations were performed on the horseshoe ( dcsc ) supercluster at the university of southern denmark . | polyamidoamine ( pamam ) dendrimers , being protonated under physiological conditions , represent a promising class of nonviral , nano - sized vectors for drug and gene delivery .
we performed extensive molecular dynamics simulations of a generic model dendrimer in a salt - free solution with dendrimer s terminal beads positively charged .
solvent molecules as well as counterions were _ explicitly _ included as interacting beads .
we find that the size of the charged dendrimer depends _ non - monotonically _ on the strength of electrostatic interactions demonstrating a maximum when the bjerrum length equals the diameter of a bead .
many other structural and dynamic characteristics of charged dendrimers are also found to follow this pattern .
we address such a behavior to the interplay between repulsive interactions of the charged terminal beads and their attractive interactions with oppositely charged counterions .
the former favors swelling at small bjerrum lengths and the latter promotes counterion condensation .
thus , counterions can have a dramatic effect on the structure and dynamics of charged dendrimers and , under certain conditions , can not be treated implicitly . |
self - interacting random processes play a prominent role in the probability theory and in statistical physic .
one fascinating aspect is that behind an apparent simplicity , they can be extremely hard to analyze rigorously .
just to mention one striking example , it is still not known whether once reinforced random walks on a ladder are recurrent in general ( see however @xcite and @xcite for a partial answer and the surveys @xcite and @xcite for other problems on reinforced processes ) .
a major difficulty in these models is that we loose the markovian property and in particular the usual dichotomy between recurrence and transience can be broken .
a famous example where this happens is for vertex reinforced random walks on @xmath1 : it is now a well known result in the field , first conjectured and partially proved by pemantle and volkov @xcite , that almost surely these processes eventually get stuck on five sites @xcite . for analogous results concerning self - attracting diffusions , see @xcite , @xcite and @xcite .
beside this very basic , yet fundamental , problem of recurrence , a question of particular interest is to understand the connections between the various discrete and continuous models . in particular
an important challenging conjecture is that self - avoiding random walks on @xmath2 converge , after renormalization , toward the @xmath3 ( see @xcite for a discussion on this and @xcite for some recent progress ) .
there are in fact not many examples where invariance principles or central limit theorems were fully established .
but for instance it was proved that random walks perturbed at extrema converge after the usual renormalization toward a perturbed brownian motion ( see e.g. @xcite and @xcite ) . in this paper
we are interested in the class of so - called excited random processes , which are among the most elementary examples of self - interacting processes . by this
we mean that the interaction with the past trajectory is as localized as possible : the evolution of these processes at any time only depend on their local time at their present position .
a discrete version was introduced relatively recently by benjamini and wilson @xcite and a generalization , called multi - excited or cookie random walks , was then further studied by zerner @xcite and many other authors ( see in particular @xcite and references therein ) . closely related models were also considered in @xcite , @xcite , @xcite and @xcite .
dolgopyat @xcite observed that in dimension @xmath4 , in the recurrent regime , and after the usual renormalization , multi - excited random walks also converge toward a perturbed brownian motion ( we will give a more precise statement later ) .
however , as we will see below , the latter are not , in some sense , the most natural continuous versions of excited processes .
somewhat more natural ones were introduced two decades ago by norris , rogers and williams @xcite , in connection with the excluded volume problem @xcite , and as a simplified model for brownian polymers .
they were later called excited brownian motions by the authors @xcite .
the aim of this paper is to show that excited brownian motions can be approached in law by multi - excited random walks in the skorokhod space , i.e. in the sense of the full process . for this
we need to use a nonstandard renormalization , namely we need to scale together and appropriately the sequence of cookies , which govern the drift of the walk , the space and the time .
now let us give more details , starting with some definitions : a multi - excited or cookie random walk @xmath5 is associated to a sequence @xmath6 of cookies in the following way : set @xmath7 for all @xmath8 , and let @xmath9 be the filtration generated by @xmath10 .
then @xmath11 and for all @xmath12 , @xmath13= 1 - { \mathbb{p}}[x_\eps(n+1)-x_\eps(n ) = -1 \mid { \mathcal{f}}_{\eps , n } ] = p_{\eps , i},\ ] ] if @xmath14 .
we notice that the case of random cookies has also been studied in the past , for instance by zerner @xcite , but here we consider only deterministic @xmath15 .
on the other hand excited brownian motions are solutions of a stochastic differential equation of the type : @xmath16 where @xmath17 is a brownian motion , @xmath18 is the local time process of @xmath19 and @xmath20 is some measurable ( bounded ) function .
so at a heuristic level the discrete and the continuous models are very similar : the drift is a function of the local time at the present position .
but the analogy can be pushed beyond this simple observation . in particular criteria for recurrence and nonzero speed in both models
( see respectively @xcite and @xcite ) are entirely similar ( see below ) .
our results here give now a concrete link .
we first prove that when @xmath0 is bounded and lipschitz , the local time process of @xmath10 , conveniently renormalized , converges to the one of @xmath19 , exactly in the same spirit as in tth s papers on self - interacting random walks ( see @xcite ) .
then we obtain a tightness result and deduce a convergence in the skorokhod space at the level of the processes ( see theorem [ cor3 ] below ) . for proving the convergence of the local time processes we use a standard criterion of ethier and kurtz @xcite on approximation of diffusions .
to show that we can apply it here we introduce some auxiliary markov chains describing the local time on each level and we make a careful analysis of the transition kernels of these markov chains ( see section [ sectheo ] ) .
_ acknowledgments : we thank an anonymous referee for encouraging us to prove the main result without the recurrence hypothesis and with weaker assumptions on the regularity of @xmath0 . _ for @xmath21 and @xmath22 , set @xmath23 consider the process @xmath24 defined by @xmath25 in particular , when @xmath26 is finite , @xmath27 .
we will say that @xmath10 is recurrent when all the @xmath26 s are finite .
a criterion for recurrence when @xmath28 for all @xmath29 or when @xmath30 for @xmath29 large enough is given in @xcite and @xcite ( namely in these cases , @xmath10 is a.s .
recurrent if , and only if , @xmath31 $ ] ) .
assume now that @xmath0 is bounded and let @xmath32 be defined by @xmath33 since @xmath0 is bounded , if @xmath34 is large enough then @xmath35 and @xmath36 is well defined .
then for @xmath37 and @xmath38 , set @xmath39,[nv]}([2nx])\quad \text{for all } x\in { \mathbb{r}}.\end{aligned}\ ] ] we give now the analogous definitions in the continuous setting .
first for @xmath37 , let @xmath40 be the right continuous inverse of the local time of @xmath19 at level @xmath41 .
again we say that @xmath19 is a.s .
recurrent if all these stopping times are a.s .
this is equivalent ( see ( * ? ? ?
* theorem 1.1 ) ) to the condition @xmath42 , where @xmath43 \ dx,\ ] ] and where @xmath44 in particular when @xmath0 is nonnegative or compactly supported this is equivalent to @xmath45 $ ] .
then set @xmath46 the ray - knight theorem describes the law of @xmath47 ( a proof is given in @xcite when @xmath48 , but it applies as well for @xmath49 ) , when @xmath50 is a.s .
finite , and we recall this result now . to fix ideas we assume that @xmath51 .
an analogous result holds for @xmath52 .
so first we have @xmath53 .
next , @xmath54 is solution of the stochastic differential equation : @xmath55,\\ \label{edsnegatif } d\lambda_{a , v}(x ) & = & 2\sqrt{\lambda_{a , v}(x)}\ , db_x + 2h\left(\lambda_{a , v}(x)\right)\ , dx \quad \hbox{for } x\in [ 0,\infty),\end{aligned}\ ] ] where @xmath17 is a brownian motion , and holds up to the first time , say @xmath56 , when it hits @xmath57 , and then is absorbed in @xmath57 ( i.e. @xmath58 for @xmath59 ) .
similarly @xmath60 is solution of ( with a drift @xmath61 instead of @xmath62 and an independent brownian motion ) up to the first time , say @xmath63 , when it hits @xmath57 , and then is absorbed in @xmath57 . for @xmath64 ,
we denote by @xmath65 the space of cdlg functions @xmath66 endowed with the usual skorokhod topology ( see for instance section 12 in @xcite ) .
the space @xmath67 will also simply be denoted by @xmath68 .
it will be implicit that all convergences in law of our processes hold in these spaces .
our first result is the following theorem : [ theo ] assume that @xmath0 is bounded and lipschitz .
assume further that for @xmath34 large enough , @xmath36 is recurrent and that @xmath19 is recurrent .
then for any finite set @xmath69 , any @xmath70 and @xmath71 , @xmath72 , @xmath73 as announced above , this theorem gives the convergence of a sequence of excited random walks toward the excited brownian motion ( associated to @xmath0 ) at the level of the local times .
we will also extend this result in a non homogeneous setting , i.e. when @xmath0 is allowed to depend also on the space variable .
we refer the reader to section [ subsecnonh ] for more details .
actually we will need this extension to prove theorem [ theo ] in the case @xmath74 .
this will be explained in section [ ige2 ] .
note that if @xmath0 is compactly supported or nonnegative ( and bounded lipschitz ) , and if @xmath75 , then @xmath19 is recurrent and @xmath36 as well for @xmath34 large enough .
a consequence of theorem [ theo ] is the following [ cor1 ] under the hypotheses of theorem [ theo ] , for any finite set @xmath69 , any @xmath70 and @xmath71 , @xmath72 , @xmath76}([nv_i])\right)_{i\in i}\quad \mathop{\longrightarrow}_{n\rightarrow\infty}^{\mathcal{l } } \quad \big(\tau_{a_i}(v_i)\big)_{i\in i}.\ ] ] for @xmath77 , @xmath72 , denote by @xmath78 some independent geometric random variables with parameter @xmath79 , independent of @xmath36 .
denote also by @xmath80 , @xmath72 , some independent exponential random variables with parameter @xmath81 , independent of @xmath19 .
then as in @xcite , we can deduce from the previous results the [ cor2 ] under the hypotheses of theorem [ theo ] , for any @xmath82 , @xmath72 , @xmath83 finally we get the following : [ cor3 ] assume that @xmath0 is bounded and lipschitz
. for @xmath84 , set @xmath85)/(2n)$ ] , which is well defined at least for @xmath34 large enough .
then @xmath86 note that , as opposed to the previous results , we do not assume in this last theorem that @xmath36 and @xmath19 are recurrent . to obtain this result
we need to prove the tightness of the sequence @xmath87)/(2n)$ ] , @xmath88 .
this is done by using a coupling between different branching processes , similar to those which were used for proving corollary [ cor1 ] .
the convergence of finite - dimensional distributions follows from corollary [ cor2 ] and an inversion of laplace transform . as for theorem [ theo ] an extension of this result to the non homogeneous setting
can be proved ( see theorem [ theofinal ] at the end of the paper ) .
let us mention now a related result of dolgopyat @xcite .
he proved a functional central limit theorem for excited random walks when @xmath15 is fixed , and in the recurrent regime ; more precisely when @xmath28 for all @xmath29 and @xmath89 . in this case
the limiting process is a perturbed brownian motion , i.e. the process defined by @xmath90 with @xmath17 a brownian motion
. we will first prove theorem [ theo ] in the case @xmath91 in section [ sectheo ] . in section [ subsecnonh ]
, we will extend the result to the non homogeneous setting and in section [ ige2 ] we will deduce the result in the general case @xmath92 .
corollaries [ cor1 ] and [ cor2 ] will be proved respectively in section [ seccor ] and [ seccor2 ] , and theorem [ cor3 ] in section [ seccor3 ] .
we assume in this section that @xmath91 .
let @xmath37 and @xmath38 be given . to fix ideas we assume that @xmath51 .
the case @xmath52 is similar .
moreover , we only prove the convergence of @xmath93 on the time interval @xmath94 , since the proofs of the convergence on @xmath95 $ ] and on @xmath96 are the same . 1 .
the sequence @xmath98 has the same law as @xmath99 , where @xmath100 is some markov chain starting from @xmath101 , which is independent of @xmath41 , 2 .
conditionally to @xmath102 , the sequence @xmath103 has the same law as some markov chain @xmath104 , starting from @xmath105 , which is independent of @xmath41 , 3 .
the sequence @xmath106 has the same law as @xmath107 , where by definition @xmath108 for all @xmath8 .
moreover , the sequences @xmath106 and @xmath103 are independent . the laws of the markov chains @xmath109 and @xmath110 will be described in subsection [ subsecmarkov ] in terms of another markov chain @xmath111 , see in particular and .
note that this idea to use the markovian property of the process @xmath112 goes back at least to kesten , kozlov and spitzer @xcite . in the following , in order to lighten the presentation we will forget about the dependence on the starting point ( which does not play any serious role here ) in the notation for @xmath113 and @xmath114 .
thus @xmath113 and @xmath115 should be understood respectively as @xmath109 and @xmath116}$ ] , where the @xmath101 will be clear from the context , and similarly for @xmath117 and @xmath118 .
now we first prove the convergence of @xmath93 on @xmath119 $ ] .
the proofs of the convergence on @xmath96 and on the full interval @xmath120 are similar and will be explained in subsection [ secainfini ] .
so on @xmath119 $ ] , @xmath93 can be decomposed as a sum of a martingale part @xmath121 and a drift part @xmath122 : @xmath123}{n}+ m^{(n)}_{a , v}(x ) + b^{(n)}_{a , v}(x ) \quad \text{for all } x\in [ a,0],\ ] ] with the following equalities in law : @xmath124-[2na ] } \big\{v_{\varepsilon_n}(k ) - { \mathbb{e}}[v_{\varepsilon_n}(k ) \mid v_{\varepsilon_n}(k-1 ) ] \big\},\end{aligned}\ ] ] and @xmath125-[2na ] } \big\{{\mathbb{e}}[v_{\varepsilon_n}(k ) \mid v_{\varepsilon_n}(k-1)]- v_{\varepsilon_n}(k-1 ) \big\}.\end{aligned}\ ] ] let also @xmath126 be the previsible compensator of @xmath127 .
we have the equality in law : @xmath128-[2na ] } \big\{{\mathbb{e}}[v_{\varepsilon_n}(k)^2 \mid v_{\varepsilon_n}(k-1)]- { \mathbb{e}}[v_{\varepsilon_n}(k)\mid v_{\varepsilon_n}(k-1)]^2 \big\},\ ] ] for all @xmath129 $ ]
. we will deduce the convergence of @xmath93 from a criterion of ethier and kurtz @xcite , namely theorem 4.1 p.354 . according to this
result the convergence on @xmath119 $ ] in theorem [ theo ] follows from propositions [ propa ] and [ propb ] below .
in addition we need to verify that the martingale problem associated with the operator @xmath130 is well posed .
this follows from theorem 2.3 p.372 in @xcite ( with the notation of @xcite take @xmath131 and @xmath132 ) .
let @xmath15 and @xmath38 be given .
we express here ( see in particular and below ) the laws of @xmath140 and @xmath141 in terms of the law of another markov chain @xmath142 .
a similar representation already appeared in tth s paper @xcite on `` true '' self - avoiding walks .
so let us first define @xmath143 by @xmath144 and for @xmath8 , @xmath145 where @xmath146 is a sequence of i.i.d random variables with uniform distribution in @xmath147 $ ] .
this @xmath148 is equal in law to the number of times the excited random walk jumps from level @xmath149 to @xmath150 , for some arbitrary @xmath151 , after @xmath29 visits at this level @xmath149 . for @xmath152 , set @xmath153 then @xmath154 is equal in law to the number of visits to level @xmath149 before the @xmath155-th jump from @xmath149 to @xmath150 .
moreover , @xmath156 is a markov chain on @xmath157 starting from @xmath57 and with transition operator @xmath158 defined for any nonnegative or bounded function @xmath159 by @xmath160 for all @xmath161 .
furthermore it is immediate that the law of @xmath162 conditionally on @xmath163 is equal to the law of @xmath164 : @xmath165 similarly the law of @xmath166 conditionally on @xmath167 is equal to the law of @xmath168 : @xmath169 by convention we denote by @xmath170 the transition operator associated to the sequence @xmath171 , where @xmath172 for all @xmath29 . in other words
@xmath173\quad \text{for all } r\in { \mathbb{n}},\ ] ] where @xmath174 is a geometric random variable with parameter @xmath175 , i.e. @xmath176 , for all @xmath177 .
note that @xmath178 and @xmath179 .
in particular , if @xmath180 is defined by @xmath181 for all @xmath161 , then for all @xmath182 , @xmath183=2m,\ ] ] where @xmath184 are i.i.d .
geometric random variables with parameter @xmath175 . note also that for all @xmath182 , @xmath185 = q_{\varepsilon}^m u ( 0)$ ] .
thus shows that @xmath186-v_{\eps}(k-1 ) = q_{\epsilon}^{v_{\eps}(k-1)}u(0)-q_0^{v_{\eps}(k-1)}u(0)+1,\ ] ] for all @xmath187 .
so in view of and , our strategy for proving proposition [ propa ] will be to estimate terms of the form @xmath188 .
note that since @xmath189 by hypothesis , we can restrict us to the case when @xmath190 .
likewise @xmath191- { \mathbb{e}}[v_{\varepsilon}(k)\mid v_{\varepsilon}(k-1)]^2= q_{\eps}^{v_{\eps}(k-1)}u^2(0 ) - ( q_{\eps}^{v_{\eps } ( k-1)}u(0))^2,\ ] ] for all @xmath187 . so in view of and we will have also to estimate terms of the form @xmath192 , for proving proposition [ propb ] . [
[ some - elementary - properties - of - the - operators - q_varepsilon - and - q_0 ] ] some elementary properties of the operators @xmath158 and @xmath170 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ for @xmath200 , let @xmath201 .
note that when @xmath202 , @xmath203 and @xmath204 are well defined for any @xmath205 .
observe also that @xmath206 , where @xmath4 is the constant function on @xmath157 .
in particular @xmath207 . moreover , for any lipschitz @xmath159 , @xmath208 , and @xmath209 . as a corollary
we get the set for all @xmath215 and @xmath177 , @xmath216 and define @xmath217 by @xmath218 this @xmath217 is a linearized version of @xmath219 , and also the first order term in the expansion of @xmath220 as @xmath221 .
the next result is immediate . by using that @xmath236 for all @xmath223 ,
we get @xmath237 thus holds with @xmath238 and @xmath239 all assertions follow immediately .
for instance we can write @xmath240 which implies ( ii ) and one can prove similarly ( i ) , ( iii ) and ( iv ) .
[ rtilde ] there exists a constant @xmath210 such that for all @xmath227 and all @xmath222 , with @xmath228 , @xmath241 where @xmath230 and @xmath231 satisfy @xmath242 where for all @xmath223 and @xmath243 , @xmath244 since @xmath245 for all @xmath223 , we get @xmath246 with @xmath247 and @xmath248 but for any @xmath223 and any @xmath177 , @xmath249 the lemma follows .
[ lemr ] there exists a constant @xmath210 such that for all @xmath227 and all @xmath222 , with @xmath228 , @xmath250 where @xmath230 and @xmath231 satisfy @xmath251 for all @xmath222 , with @xmath228 , and all @xmath252 , @xmath253 where @xmath254 and @xmath255 satisfy just recall that for all @xmath29 and @xmath223 , @xmath261 $ ] , where @xmath262 are i.i.d .
geometric random variables with parameter @xmath175 .
thus , @xmath263 , where @xmath264 + { \mathbb{e}}[(\xi_1+\dots+\xi_i)g(r+\xi_1+\dots+\xi_i)],\ ] ] and @xmath265.\ ] ] claims ( i ) , ( ii ) , ( iii ) and ( iv ) follow immediately frome these expressions ( by using also that @xmath266=2i$ ] ) . next write @xmath267\\ & \le & \sqrt{2i } \textrm{lip}(f),\end{aligned}\ ] ] by using cauchy - schwarz inequality and the fact that @xmath268 and @xmath269 , for all @xmath29 .
we also have @xmath270\\ & \le & \sqrt{2i}\left(|g|_\infty + ( r+2i ) \textrm{lip}(g)\right).\end{aligned}\ ] ] we have latexmath:[\ ] ] on the other hand corollary [ cor2 ] shows that the term on the left hand side converges toward @xmath547 $ ] .
since this holds for any @xmath377 and any @xmath501 , we deduce that @xmath548 and @xmath549 have the same law for every @xmath84 ( see ( * ? ? ? * theorem 1a p.432 ) ) .
this proves the convergence of one - dimensional distributions .
this finishes the proof under the additional hypothesis of recurrence and it just remains to explain how one can remove this hypothesis . for this
we use a truncation argument . for
any @xmath133 and @xmath88 , let @xmath550 , resp .
@xmath551 , be the functions equal to @xmath552 , resp .
@xmath553 , when @xmath554 , resp . when @xmath555 , and equal to zero otherwise .
it is immediate that @xmath556 still converges to @xmath557 as in .
denote now by @xmath558 and @xmath559 the processes associated respectively to @xmath556 and @xmath557 . since @xmath0 and the @xmath560 s
are bounded , and since they are equal to zero outside of @xmath561 , these processes are recurrent .
so we just have proved that @xmath558 converges in law toward @xmath559 .
note now that up to the time @xmath562}(0)\wedge \tau_{\eps_n,[-2nr]}(0))/(4n^2)$ ] , ( with the notation from section [ seccor2 ] ) , @xmath558 and @xmath532 are equal in law .
similarly up to the time @xmath563 , @xmath559 and @xmath19 are equal in law . but
corollary [ cor1 ] shows that @xmath564}(0)/(4n^2)$ ] and @xmath565}(0)/(4n^2)$ ] , converge in law respectively toward @xmath566 and @xmath567 . by using also the monotonicity in @xmath138 of these random times
, we deduce that for any @xmath368 , @xmath568}(0)\wedge \tau_{\eps_n,[-2nr]}(0 ) \le 4n^2t\right ) \to 0,\ ] ] when @xmath569 , uniformly in @xmath34 .
it follows immediately that for any @xmath368 , @xmath532 converges in law to @xmath19 on the time interval @xmath570 $ ] .
since this is true for any @xmath368 , this concludes the proof of theorem [ theofinal ] . _
the notation @xmath571 is taken from tth and werner @xcite .
we notice by the way that here also the set @xmath572 forms a family of reflected / absorbed coalescing processes . in @xcite
the @xmath54 s were moreover independent brownian motions ( reflected or absorbed in @xmath57 depending on the time interval ) and therefore @xmath573 was called ( in their section 2.1 ) a ficrab ( for family of independent coalescing reflected and absorbed brownian motions ) .
such family of coalescing brownian motions seems to have been first studied by arratia @xcite and is now better known under the name of brownian web ( see for instance @xcite ) .
here the situation is slightly different : first each @xmath54 is some diffusion which is not a brownian motion and before they coalesce two @xmath54 s are not independent .
for instance if @xmath574 , then @xmath575 satisfies the following system of stochastic differential equations : @xmath576 for all @xmath577 , where @xmath17 and @xmath578 are two independent brownian motions .
this result follows from and the ray - knight theorem ( see for instance ( * ? ? ?
* theorem 6.1 ) ) .
note that we could describe similarly the law of @xmath579 , for any finite set @xmath69 , and any @xmath580 , @xmath72 . in @xcite ,
the family @xmath573 was called a sequence of forward lines and the dual sequence , the sequence of backward lines @xmath581 , was defined by @xmath582 for all @xmath583 and @xmath38 . as in @xcite
we can define @xmath584 here and we have also @xmath585 in law ( see theorem 2.3 in @xcite ) .
it is important to observe that @xmath586 we notice now some other notable differences with the situation in @xcite .
first if we denote by @xmath587 the law of @xmath573 , then the law of @xmath584 is @xmath588 .
in particular @xmath573 and @xmath584 do not have the same law ( in other words @xmath573 is not self - dual ) , except if @xmath589 .
moreover , for any @xmath41 ( say @xmath590 ) and @xmath38 , the process @xmath54 will almost surely not hit @xmath57 in the time interval @xmath119 $ ] .
the reason is that in the time interval @xmath591 $ ] the excited bm will cross each level @xmath129 $ ] and strictly increase its local time on these levels ( by using the absolute continuity between the laws of a standard bm and the excited bm ) .
similarly given any @xmath592 , @xmath101 and @xmath420 , we have @xmath593 for all @xmath594 $ ] almost surely .
let us also notice that couples of processes such as @xmath595 and @xmath596 , if @xmath590 , are _ conjugate diffusions _ ( see @xcite for a definition ) .
similarly @xmath597 and @xmath598 , if @xmath590 , are also conjugate .
+ now we can sketch another proof of theorem [ cvlawfd ] which bypass the use of corollary [ corext ] and uses instead these notions of forward and backward lines .
the idea is to first prove that @xmath599 this can be done by using ethier - kurtz s result ( theorem 4.1 p.354 in @xcite ) , and .
one can next define analogues @xmath600 and @xmath601 respectively of @xmath573 and @xmath584 , in the discrete setting and it then suffices to use ( and its discrete counterpart ) , and to deduce the desired convergence .
since we already gave another proof , we omit the details here . _
* billingsley p. : * _ convergence of probability measures _
, second edition .
wiley series in probability and statistics : probability and statistics . a wiley - interscience publication .
john wiley & sons , inc .
, new york , ( 1999 ) , x+277 pp . * lawler g. , schramm o. , werner w. : * _ on the scaling limit of planar self - avoiding walk _ ,
fractal geometry and application , a jubilee of benoit mandelbrot , proc .
pure math .
72 , part 2 , amer .
soc . , providence , ri , ( 2004 ) , 339364 . | we obtain the convergence in law of a sequence of excited ( also called cookies ) random walks toward an excited brownian motion .
this last process is a continuous semi - martingale whose drift is a function , say @xmath0 , of its local time .
it was introduced by norris , rogers and williams as a simplified version of brownian polymers , and then recently further studied by the authors . to get our results we need to renormalize together the sequence of cookies , the time and the space in a convenient way .
the proof follows a general approach already taken by tth and his coauthors in multiple occasions , which goes through ray - knight type results .
namely we first prove , when @xmath0 is bounded and lipschitz , that the convergence holds at the level of the local time processes .
this is done via a careful study of the transition kernel of an auxiliary markov chain which describes the local time at a given level .
then we prove a tightness result and deduce the convergence at the level of the full processes . |
transfer operators can be regarded as global representations of a system s dynamics .
their spectral properties yield insight into the statistical long - term behaviour of the underlying system , such as rates of mixing or the existence of central limit theorems ( see , for example , @xcite ) .
if @xmath4 is the set of local inverse branches of a real analytic expanding map , then the associated transfer operator @xmath5 , defined by @xmath6 preserves and acts compactly on certain spaces of holomorphic functions ( see , for example , @xcite ) .
is referred to as the perron - frobenius operator .
more general transfer operators can be obtained by replacing the @xmath7 in by other suitable weight functions . ] in particular , its spectrum is a sequence of eigenvalues @xmath8 converging to zero , together with zero itself . assuming that the eigenvalues of @xmath9 are ordered with respect to decreasing modulus , the second eigenvalue @xmath10 determines the exponential rate of mixing for generic analytic observables . a faster exponential rate of mixing can occur if one chooses ( non - generic ) observables which do not ` feel ' the rate corresponding to the first @xmath11 subleading eigenvalues , that is , observables in a subspace complementary to the eigendirections of @xmath12 .
consequently , knowledge of the whole spectrum @xmath13 is useful , since it determines the set of all possible exponential mixing rates , also known as the correlation spectrum in this context . surprisingly there
exist only few examples of maps in the literature for which the spectrum of the corresponding transfer operator is known explicitly .
the only one - dimensional examples known to the authors are piecewise linear interval maps , or more generally , piecewise linear markov interval maps ( see , for example , @xcite ) and circle maps of the form @xmath14 for @xmath11 a non - zero integer .
it turns out that for the latter maps the spectra of the corresponding transfer operators when acting on analytic functions are all identical and coincide with the two - point set @xmath15
. see ( * ? ? ?
* exercise 2.15 ) for a proof when @xmath16 ; the general case can be proved along the same lines . as a result
, the question arises as to whether there are analytic circle maps for which the associated transfer operator has infinitely many non - zero eigenvalues .
the purpose of this paper is twofold : firstly , to construct examples of circle maps with infinitely many explicit non - zero eigenvalues in the spectrum of @xmath5 . denoting the circle by @xmath17 , and letting @xmath18 be the space of bounded holomorphic functions on an annulus @xmath19 containing @xmath20 ( see section [ sec : rigorous ] for precise definitions ) ,
we can state the following .
[ thm : main ] for any @xmath0 with @xmath1 there exists an analytic expanding circle map @xmath21 such that the eigenvalues of the associated transfer operator @xmath5 in , when acting on @xmath18 , are precisely all nonnegative powers of @xmath2 and @xmath3 , that is , the spectrum of @xmath5 is @xmath22 moreover , the algebraic multiplicity of the leading eigenvalue is one , while the algebraic multiplicity of the remaining eigenvalues is two for real @xmath2 and one for @xmath2 with nonvanishing imaginary part .
secondly , the analytic maps arising from theorem [ thm : main ] yield interval maps which provide counterexamples to the following conjecture .
[ conj : levinmayer ] let @xmath23 be a bounded domain with @xmath24 and @xmath25 contracting holomorphic mappings with their unique fixed points @xmath26 in @xmath27 .
if the @xmath28 are real , then all eigenvalues of the corresponding transfer operator @xmath29 with small enough modulus are real .
mayer @xcite originally conjectured that transfer operators satisfying the hypotheses of the above conjecture have real spectra .
counterexamples to mayer s conjecture were given by levin in @xcite which led to the above weakening of mayer s conjecture . before proceeding to definitions and rigorous proofs
, we shall briefly explain the genesis of our result . turning the classical question of finding eigenvalues of @xmath30 for a given map on its head
, we attempt to construct a map for which the transfer operator has a given eigenvalue and a given eigenfunction . considering an analytic expanding map on the interval @xmath31 with two full branches ,
the eigenvalue problem of the transfer operator formally reads @xmath32 where @xmath33 and @xmath34 are an eigenvalue and eigenfunction of @xmath30 .
given an analytic invariant density @xmath35 one may consider for @xmath36 with @xmath37 and @xmath38 as an equation to compute suitable inverse branches @xmath39 and @xmath40 of the map .
this setup is a particularly simple case of the so called inverse perron - frobenius problem @xcite , which has been applied in various guises to taylor - make chaotic maps with given stationary properties ( see , for example , @xcite ) . as we are attempting to construct a map with two branches we are at liberty to specify a nontrivial eigenvalue and corresponding eigenfunction .
thus , given an invariant density @xmath35 , a real number @xmath2 with @xmath1 , and a potential eigenfunction @xmath41 , we seek to solve @xmath42 for the inverse branches @xmath39 and @xmath40 . here
p and u denote suitable antiderivatives of @xmath35 and @xmath41 , respectively . a priori
, there is no guarantee that admits a real solution for @xmath39 and @xmath40 and that such a solution actually determines an analytic full branch interval map . developing conditions under which this is the case seems to be a challenging task .
nevertheless , if we fix the interval @xmath43 $ ] , take @xmath35 to be the uniform density , and @xmath44 for the eigenfunction with eigenvalue @xmath2 , then leads to @xmath45 after a short calculation it is possible to deduce from an explicit expression for @xmath39 and @xmath40 , and , as we shall see in section [ sec : circle ] , this indeed determines an analytic full branch map .
it turns out that for this particular example the complete spectrum can be obtained , an observation that provides the main content of theorem [ thm : main ] .
certainly , this reasoning does not generally provide examples with fully prescribed spectrum .
however , the method deserves further exploration , as it furnishes maps with given partial spectral information .
the paper is structured as follows . in section [ sec : rigorous ] we introduce analytic expanding circle maps , and define banach spaces of holomorphic functions on which the associated transfer operators @xmath5 are compact .
section [ sec : circle ] is devoted to the construction of a family of circle maps and the proof of theorem [ thm : main ] . in section [ sec : interval ]
we consider the same family of maps on an interval and thus obtain counterexamples to conjecture [ conj : levinmayer ] .
the main purpose of this section is to define suitable function spaces on which transfer operators induced by analytic expanding circle maps are compact .
we start by defining what is meant by an analytic expanding circle map .
we say that @xmath46 is an _ analytic expanding circle map _ if the following two conditions hold : a. @xmath21 is analytic on @xmath20 ; b. @xmath47 .
it is not difficult to see that @xmath21 is a @xmath48-fold covering of @xmath49 for some integer @xmath50 .
moreover , the map @xmath21 has analytic extensions to certain annuli containing @xmath49 . with slight abuse of notation we shall write @xmath21 for the various extensions as well . to be precise , for @xmath51 let @xmath52 denote the open annulus @xmath53 and write @xmath54 the expansivity of @xmath21 yields the following result .
[ lem : expand ] if @xmath21 is an analytic expanding circle map , then there is @xmath55 such that a. both @xmath21 and @xmath56 are analytic on the closure @xmath57 of @xmath58 ; b. @xmath59 , where @xmath60 denotes the boundary of @xmath58 . since @xmath21 is an analytic expanding circle map it is possible to choose @xmath61 such that both @xmath21 and @xmath56 are analytic on @xmath62 with @xmath63 it is not difficult to see that @xmath64 is differentiable for all @xmath65 with @xmath66 and @xmath67 where @xmath68 and @xmath69 denote the real and imaginary part of @xmath70 .
since @xmath21 leaves @xmath49 invariant , equation ( [ e : phide ] ) implies either @xmath71 or @xmath72 suppose now that ( [ e : pos ] ) holds ( the other case can be dealt with similarly ) .
fixing @xmath73 with @xmath74 we can choose @xmath75 with @xmath76 and @xmath77 , @xmath78 such that @xmath79 , \forall \theta \in \mathbb r)\,.\ ] ] equation ( [ e : rde ] ) now implies @xmath80 and @xmath81 thus @xmath82 so @xmath83 has all the desired properties . given an expanding circle map @xmath21 , we associate with it a transfer operator @xmath30 by setting @xmath84 where @xmath85 to denotes the @xmath86-th local inverse of @xmath21 .
it turns out that @xmath9 is well - defined and bounded as an operator on @xmath87 .
although this is a standard result , we shall provide a short proof , since part of the argument will play a crucial role later on . in the following
we shall use @xmath88 to denote a simple closed positively oriented path along the circle centred at the origin with radius @xmath35 .
[ lem : dual ] for any @xmath89 and any @xmath90 we have @xmath91 in particular , the transfer operator @xmath30 is bounded as an operator from @xmath87 to @xmath87 .
using change of variables we see that @xmath92 where we have used the fact that @xmath93 up to a set of measure zero .
finally , the assertion that @xmath30 maps @xmath87 continuously into itself follows from equation ( [ lem : dual : e ] ) , since for any @xmath89 and any @xmath90 @xmath94 it turns out that @xmath9 leaves certain subspaces of @xmath95 invariant . of particular interest
are spaces consisting of holomorphic functions . for @xmath96 , an open subset of @xmath97 , we write @xmath98 to denote the banach space of bounded holomorphic functions on @xmath96 equipped with the norm @xmath99 .
the proof of the invariance of @xmath100 under @xmath9 for suitable @xmath96 will rely on fourier theory . here and in the following , we shall use @xmath101 to denote the @xmath11-th fourier coefficient of @xmath102 . before stating the next result we require some more notation . given two subsets @xmath96 and @xmath103 of @xmath104
we write @xmath105 if @xmath106 is a compact subset of @xmath103 .
we now have the following result .
[ lem : ltilde ] suppose that annuli @xmath19 and @xmath107 in @xmath108 have been chosen .
] such that @xmath109 then the transfer operator @xmath30 maps @xmath110 continuously to @xmath111 .
given @xmath112 , we shall show that @xmath113 by estimating the asymptotic behaviour of the fourier coefficients of @xmath114 .
write @xmath115 and @xmath116 to denote the radii of the circles forming the ` exterior ' boundary of @xmath58 and @xmath19 , respectively ( see figure [ fig : annuli ] ) .
next choose @xmath117 with @xmath118 similarly , write @xmath119 and @xmath120 to denote the radii of the circles forming the ` interior ' boundary of @xmath58 and @xmath19 , respectively and choose @xmath121 with @xmath122 fix @xmath112 with @xmath123 and let @xmath124 .
using lemma [ lem : dual ] we see that @xmath125 similarly , for @xmath126 we have @xmath127 hence , @xmath128 converges absolutely for all @xmath129 and @xmath130 converges absolutely for all @xmath131 .
moreover , for @xmath132 we have @xmath133 thus , by the uniqueness of the fourier transform on @xmath87 , we conclude that @xmath134 and @xmath135 choosing @xmath136 in the previous lemma shows that @xmath30 induces a well defined continuous operator from @xmath111 to itself .
it turns out that @xmath137 is compact . in order to prove this result
, we shall employ a factorisation argument . given @xmath138 with @xmath139
define the canonical embedding @xmath140 @xmath141 the embedding @xmath142 is compact by montel s theorem ( see , for example , ( * ? ? ?
* chapter 7 , theorem 2.9 ) ) and , as we shall shortly see , is well - approximated by the following operators : for @xmath143 a positive integer , define the finite rank operator @xmath144 by @xmath145 the approximability of @xmath142 alluded to earlier is the content of the following result .
[ lem : j ] let @xmath142 and @xmath146 be defined as above .
then @xmath147 in particular , the embedding @xmath142 is compact .
choose @xmath148 with @xmath149 let @xmath150 and @xmath151 denote the radii of the circle forming the ` exterior ' and ` interior ' boundary of @xmath107 , respectively , and let @xmath152 and @xmath153 denote the oriented ` exterior ' and ` interior ' boundary of @xmath154 , respectively , so that @xmath155 fix @xmath156 with @xmath157 .
then @xmath158 from which the assertions follow .
now , the transfer operator @xmath159 , factorises as @xmath160 where @xmath161 is the canonical embedding , which is compact by lemma [ lem : j ] , and @xmath162 is the transfer operator viewed as an operator from @xmath110 to @xmath111 , guaranteed to be continuous by lemma [ lem : ltilde ] .
thus , the factorisation ( [ eq : l ] ) implies the following result .
[ cor : compact ] let @xmath163 with @xmath164 .
then @xmath165 is compact .
in this section we introduce a family of analytic expanding circle maps for which we are able to explicitly determine the spectrum of @xmath166 as already mentioned in the introduction , the family arises from the construction of an expanding map for which the transfer operator has a specified eigenfunction for a given eigenvalue @xmath2 .
a short calculation reveals that for @xmath167 a solution to lifts to @xmath168 , where @xmath169 this is an increasing diffeomorphism with inverse @xmath170 given by @xmath171 note that @xmath172 is a lift of a circle map @xmath46 , which satisfies @xmath173 and @xmath174 where @xmath175 is the projection map defined by @xmath176 the map @xmath21 is a twofold covering of @xmath177 ( see figure [ fig : map ] ) .
note that @xmath178 for @xmath167 .
thus @xmath21 is an analytic expanding circle map .
it turns out that @xmath21 can be written in closed form . using the relation @xmath179 ,
it follows that , for @xmath180 , @xmath181 which gives @xmath182 it is not difficult to see that the above expression for @xmath21 yields an analytic expanding circle map not just for real @xmath2 , but for any @xmath0 with @xmath1 ( see figure [ fig : map ] ) .
it is possible to write down lifts of ( [ eq : tau ] ) for complex @xmath2 similar to .
in fact , a short calculation shows that if @xmath183 then the argument of @xmath184 in needs to be replaced by @xmath185 . given @xmath21 as in ( [ eq : tau ] ) , we now choose an annulus @xmath186 with @xmath164 . by proposition
[ cor : compact ] the associated transfer operator @xmath187 is well - defined and compact . as we shall see , the spectrum of @xmath5 can be computed by analysing the spectrum of a suitable matrix representation , which is obtained as follows . for @xmath188
consider the projection @xmath189 given by the same functional expression as @xmath190 in , now viewed as an operator from @xmath111 to @xmath111 .
clearly , @xmath191 is an operator of rank @xmath192 .
writing @xmath193 , the set @xmath194 is a basis for @xmath195 and the restriction of @xmath196 to @xmath197 is represented by the @xmath198 matrix @xmath199 defined by @xmath200 in particular , the non - zero spectrum of @xmath196 is given by the non - zero spectrum of @xmath199 .
observe that defines an infinite matrix @xmath201 containing @xmath199 as a finite submatrix .
the following lemma summarizes the properties of @xmath201 .
[ lem : l_explicit ] for @xmath202 the following hold : a. @xmath203 ; b. @xmath204 if @xmath205 ; c. @xmath206 ; d. @xmath207 for @xmath208 ; e. @xmath209 for @xmath210 .
assertions ( a ) and ( b ) immediately follow from , + while ( c ) is a consequence of @xmath211 for ( d ) and ( e ) , observe that @xmath212 is holomorphic for all @xmath213 in the closed unit disk .
thus , by the residue theorem , @xmath214 finally @xmath210 implies @xmath215 , as the intergrand is a holomorphic function .
the lemma above implies that @xmath199 has the following upper - lower triangular matrix structure and @xmath216 we have @xmath217 ] @xmath218 clearly , the spectrum of @xmath199 is given by the diagonal elements @xmath219 , that is , @xmath220 moreover , the triangular structure of @xmath199 implies @xmath221 . before embarking on the proof of our main result
, we require one more lemma , which relates the eigenvalues of @xmath30 with the eigenvalues of @xmath199 .
[ cor : eigenval ] let @xmath163 with @xmath164 and suppose that @xmath222 for every @xmath188 .
then the non - zero eigenvalues ( with multiplicities ) of @xmath223 are precisely the non - zero eigenvalues of @xmath199 as @xmath143 tends to infinity . in particular , the spectrum of @xmath30 is given by @xmath224 clearly @xmath225 .
since @xmath222 we have @xmath226 for every @xmath188 . using the factorisation
we see that @xmath227 which , using lemmas [ lem : ltilde ] and [ lem : j ] , implies @xmath228 this , together with ( * ? ? ?
* xi.9.5 ) guarantees that every non - zero eigenvalue of @xmath30 is an eigenvalue of @xmath199 for some @xmath229 .
we are now able to prove our main result . for every @xmath230 with @xmath231 , the map @xmath21 in is an analytic expanding circle map . by lemma [ cor : eigenval ] the spectrum of the associated @xmath232 consists of eigenvalues , together with zero , given by @xmath233 the assertions concerning the multiplicities of the eigenvalues of @xmath30 follow from the corresponding properties of @xmath199 .
in the previous section we have considered the transfer operator @xmath234 associated to an analytic expanding circle map @xmath235 which maps the space of bounded holomorphic functions on an appropriately chosen annulus @xmath236 compactly into itself .
the circle map @xmath21 gives rise to a map @xmath237 on an interval @xmath238 $ ] , chosen such that a fixed point @xmath239 of @xmath21 corresponds to the interval endpoint @xmath240 .
choosing a suitable complex neighbourhood @xmath241 of @xmath31 , we shall now study the spectral properties of @xmath242 , the transfer operator corresponding to @xmath237 .
more precisely , let @xmath243 denote the interval map arising from the circle map @xmath244 via @xmath245 with a projection @xmath246 satisfying @xmath247 .. ] let @xmath248 be the set of inverse branches of @xmath237 . with slight abuse of notation we keep writing @xmath237 and @xmath249 for the respective analytic extensions to neighborhoods containing @xmath31 .
since @xmath21 is an analytic k - covering , we have the matching conditions ( with suitable labelling of the inverse branches ) @xmath250 where for each @xmath251 , we use @xmath252 to denote the @xmath11-th derivative of @xmath249 .
since @xmath237 is expanding , all inverse branches @xmath249 are contractions .
we can thus choose a topological disk @xmath241 containing @xmath31 such that @xmath253 and @xmath254 for all @xmath86 .
then @xmath255 , given by @xmath256 yields a bounded operator . moreover , @xmath257 is compact ( see , for example , @xcite ) , its spectrum consisting of countably many eigenvalues accumulating at zero only .
[ rem : fp ] it is perhaps not surprising that the operators @xmath258 and @xmath259 are closely related . in order to see this
, we define the operator @xmath260 by @xmath261 clearly @xmath262 implies @xmath263 injective . however , the operator @xmath263 is not surjective , as the image @xmath264 is not all of @xmath265 .
it is easy to verify that @xmath266 and @xmath267 are related by @xmath268 and that @xmath269 , which follows from the injectivity of @xmath263 . on the other hand ,
an eigenvalue of @xmath270 with an eigenfunction @xmath271 is also an eigenvalue of @xmath272 if @xmath273 .
the following lemma connects the spectrum of @xmath257 with the spectrum of @xmath274 .
this result is mentioned in the introduction of @xcite together with a proof based on the theory of fredholm determinants . here
we shall give a short alternative proof .
[ lem : interval ] suppose that @xmath21 is an analytic expanding circle map and @xmath243 the corresponding interval map fixing the interval endpoint @xmath240 .
let @xmath274 and @xmath275 be the corresponding transfer operators as defined above .
then the spectrum of @xmath257 is given by @xmath276 let @xmath277 denote the strong dual of @xmath265 and let @xmath278 denote the dual operator of @xmath257 , that is , @xmath279 for @xmath280 , let @xmath281 be defined by @xmath282 it is not difficult to see that @xmath283 is an eigenvector of @xmath284 with eigenvalue @xmath285 since @xmath286 where the penultimate equality follows from .
we can proceed similarly for an arbitrary @xmath251 .
observe that the @xmath11-th derivative of @xmath287 is given by @xmath288 where each @xmath289 is a weight function composed of derivatives of @xmath249 of order up to @xmath290 , and in analogy with satisfying @xmath291 for @xmath292 . a calculation similar to the above yields @xmath293 it follows that @xmath294 , where @xmath295 for each @xmath11 .
thus @xmath296 is an eigenvalue of @xmath297 and hence of @xmath275 .
as @xmath298 and every eigenvalue of @xmath274 is an eigenvalue of @xmath257 , we have shown @xmath299 for the converse inclusion recall remark [ rem : fp ] and assume that @xmath300 is an eigenfunction of @xmath257 with eigenvalue @xmath301 and @xmath302 .
it follows that there is @xmath303 such that @xmath304 and @xmath305 for @xmath306 , from which @xmath307 for @xmath306 .
since @xmath308 , this implies @xmath309 as @xmath310 is linear and non - zero , it follows that @xmath311 .
we can now apply this result to the interval maps arising from the family of circle maps defined in section [ sec : circle ] .
let @xmath43 $ ] and @xmath312 with @xmath1 , then the interval map @xmath237 arising from @xmath21 in fixes the interval endpoint @xmath313 with @xmath314 . by theorem
[ thm : main ] and lemma [ lem : interval ] , the eigenvalues of @xmath259 can be divided into two classes , those given by the eigenvalues of @xmath258 ( each of multiplicity two , except the eigenvalue @xmath315 of multiplicity one ) and those by the powers of the inverse multiplier of the fixed point @xmath240 , that is , @xmath316 see also figure [ fig : specint ] .
considering @xmath317 , say @xmath318 with @xmath1 , we now obtain counterexamples to conjecture [ conj : levinmayer ] .
the fixed point of @xmath21 is @xmath319 with @xmath320 as above , the spectrum of @xmath259 splits into two parts : @xmath321 note that the transfer operator @xmath257 associated to @xmath237 satisfies the conditions of conjecture [ conj : levinmayer ] , but , for @xmath322 , has countably infinitely many non - real eigenvalues of arbitrarily small modulus . to the best of our knowledge
these are the first examples of nontrivial circle and interval maps for which the entire spectrum of the associated perron - frobenius operators is known explicitly .
certain conjectures were previously hard to test , without examples .
these might now be more accessible .
we would like to thank hans - henrik rugh for pointing out to us that transfer operators of circle maps require a holistic approach .
w.j . gratefully acknowledges support by epsrc ( grant no .
ep / h04812x/1 ) . | we show that for any @xmath0 with @xmath1 there exists an analytic expanding circle map such that the eigenvalues of the associated transfer operator ( acting on holomorphic functions ) are precisely the nonnegative powers of @xmath2 and @xmath3 . as a consequence
we obtain a counterexample to a variant of a conjecture of mayer on the reality of spectra of transfer operators . |
finding galaxies at large redshifts is important for our understanding of galaxy formation in the early universe .
selecting ultra steep spectrum radio sources ( @xmath9 , see sect .
2 ) has proven to be the most efficient way of finding distant galaxies ( e.g. chambers et al .
1990 ; mccarthy 1993 ) .
searches based on this method have resulted in the discovery of many high redshift ( @xmath10 ) radio galaxies ( hzrgs ) , but few with @xmath11 ( @xcite ) . in the last few years , we have carried out a programme to enlarge the sample of hzrgs , by selecting ultra steep spectrum ( uss ) radio sources ( e.g. rttgering 1993 ) .
this project has included radio imaging ( @xcite ) , optical broadband imaging ( @xcite ) and spectroscopy of potential distant galaxies .
the optical observations were carried out as a key programme at the european southern observatory in chile .
our uss radio source survey has resulted in the discovery of more than 30 radio galaxies at redshifts @xmath10 ( @xcite ) .
we have now started follow up observations to study individual objects in more detail .
this includes high resolution spectroscopy and narrow band imaging of the ly@xmath2 emitting gas , infrared spectroscopy and a sensitive high resolution radio survey of the objects . here
we report on the most distant galaxy discovered in the eso key programme , 1243 + 036 (= 4c03.24 ) at z=3.57 .
this is one of the most distant galaxies reported until now and its redshift is only exceeded by 4c41.17 at @xmath12 ( @xcite ) and 8c1435 + 63 at @xmath13 ( @xcite ) . in sect . 2 we describe the selection of the object , the observations and data analysis .
in sect .
3 the observational results are presented and physical parameters are deduced .
subsections describe the determination of the redshift and the kinematics of the ly@xmath2 gas ( 3.1 ) , the properties of the radio source ( 3.2 ) , the morphology of the optical continuum and ly@xmath2 emission ( 3.3 ) and a summary of the various radio and optical components ( 3.4 ) . in sect .
4 we discuss the observed characteristics of 1243 + 036 and in particular the kinematics of the ionized gas .
we discuss the ionization and density of the ly@xmath2 gas ( 4.1 ) , the kinematics of the ionized gas and scenarios for the origin of the outer halo ( 4.2 ) , the bent radio structure and its relation to the ionized gas ( 4.3 ) and the aligned optical continuum emission ( 4.4 ) .
we summarize our conclusions in sect . 5 .
throughout this paper we assume a hubble constant of @xmath14 km s@xmath5 mpc@xmath5 and a deceleration parameter of @xmath15 .
as a basis for our survey to find distant galaxies , several radio catalogues at different frequencies and different overlapping regions of the sky were combined to select samples of uss ( radio spectral index @xmath9 , where s@xmath16 @xmath17 @xmath18 , with s@xmath16 the flux density and @xmath19 the frequency ) radio sources ( @xcite ) .
the samples were selected at a range of frequencies from 38 mhz to 408 mhz .
the radio galaxy 1243 + 036 is part of the 178 mhz selected sample ( @xcite ) , and is also known as 4c03.24 .
the samples of uss sources were then observed with the vla as a preliminary for optical imaging and spectroscopy ( @xcite ) . the @xmath20 resolution vla image of 1243 + 036 from rttgering et al .
( 1994 ) at 1465 mhz shows a simple , although bent , double morphology , with a full angular extent of 8@xmath21 .
the source has a total flux density at 1.5 ghz of 310 mjy , and has a very steep integrated radio spectrum , with a spectral index , @xmath2 , between 178 mhz and 2700 mhz of @xmath22 .
1243 + 036 was optically identified from @xmath1-band ccd imaging of these samples of radio sources using the eso - mpi 2.2 m telescope .
these observations are described in rttgering et al .
( 1995b ) . here
we describe low resolution spectroscopic observations to determine the redshift of 1243 + 036 and follow up observations using high resolution spectroscopy and imaging of the ly@xmath2 gas and the radio source to study the radio galaxy in more detail .
a summary of the various observations is given in table 1 .
low resolution optical spectroscopy was carried out on the eso new technology telescope ( ntt ) , using the eso multi - mode instrument ( emmi ) .
the detector was a thomson ccd with 1024@xmath23 pixels that has a scale along the slit of @xmath24 per pixel .
the spectrum covered the wavelength region from 4100 to 7900 .
a @xmath25 wide slit was used , oriented along the radio axis ( position angle 152@xmath26 , pa1 in table 1 ) , giving a spectral resolution of 12 ( fwhm ) .
four separate exposures of 30 minutes each were taken on 21 march 1991 .
the seeing during the observations was @xmath27 .
the conditions were photometric and the spectroscopic standard star eg54 was observed for flux calibration ( @xcite ) .
the raw spectra were flat - fielded , sky - subtracted , flux - calibrated and corrected for atmospheric extinction using the longslit package in the iraf reduction package of the u.s .
national optical astronomy observatory .
the accuracy of the wavelength calibration was @xmath83 and we estimate that the flux calibration is good to @xmath810% .
a one - dimensional spectrum of 1243 + 036 was extracted using an aperture of @xmath28 to include all emission from the detected spatially extended emission line ( see sect .
high resolution optical spectroscopy of the ly@xmath2 emission line was carried out with emmi on the eso ntt on 14 april 1994 .
the detector was a tektronix ccd having 2048@xmath23 pixels with a scale along the slit of @xmath29 per pixel .
the ccd was binned 2x2 giving an effective scale along the slit of @xmath30 per pixel . using a @xmath31 wide slit with eso grating 6 ,
the spectral resolution was 2.8 .
four 1 hour exposures were taken with a slit orientation along the radio axis ( 152@xmath26 , pa1 ) , and one 1 hour exposure with the slit orientation angle orthogonal to the radio axis ( at 62@xmath26 , pa2 ) .
conditions were photometric and the seeing was approximately @xmath32 .
the spectra were reduced in the same manner as the low resolution spectra described above .
the skyline at 5577 leaves strong residuals after sky subtraction and this region of the spectrum was therefore not taken into account in the analysis . after the initial vla `` snapshot '' observations described in rttgering et al .
( 1994 ) , further radio observations of 1243 + 036 were made on august 20 , 1991 and march 18 , 1994 , using the vla in ` a ' , configuration ( see table 1 ) .
the march 18 , 1994 observations were part of a high resolution survey of a large sample of powerful radio galaxies at redshifts @xmath33 ( @xcite ) .
all observations were made using a 50 mhz bandwidth .
data processing was performed using the astronomical image processing system ( aips ) .
the system gains were calibrated with respect to the standard sources 3c48 and 3c286 .
we estimate that the flux scale is accurate to better than @xmath34% at all frequencies .
phase calibration was performed using the nearby calibrator 1236 + 077 .
the ( on - axis ) antenna polarization response terms were determined , and corrected - for , using multiple scans of the calibrator 0746 + 483 over a large range in parallactic angle .
absolute linear polarization position angles were measured using two scans of 3c286 separated in time by a few hours . from the difference in solutions between these two we estimate the systematic uncertainty in the observed polarization position angles to be @xmath81@xmath26 at all frequencies .
the calibrated data were then edited and self - calibrated using standard procedures to improve image dynamic range .
we have also combined the 1.5 ghz data presented in rttgering et al .
( 1994 ) with the new data from august 20 , 1991 .
the gridded visibilities were uniformly weighted to maximize spatial resolution .
images of the three stokes polarization parameters , i , q and u were synthesized , and all images were cleaned down to the level of 2.5 times the theoretical rms on the image , using the algorithm as implemented in the aips task mx . the observations at the different frequencies were added in the image plane , weighted by the root rms of the images , to produce the final maps at 8.3 ghz , 4.7 ghz and 1.5 ghz .
narrow band imaging was carried out with the eso ntt using the superb seeing imager ( susi ) .
a narrow band filter ( eso filter # 731 ) was used that has a central wavelength of 5571 and a bandpass width of 60 ( fwhm ) .
the detector was a tektronix ccd with 1024@xmath23 pixels and a scale of @xmath35 per pixel .
four separate exposures of 30 minutes were taken each with a seeing better than @xmath7 .
each image was shifted by about @xmath36 with respect to the previous to minimize problems due to flat - fielding and to facilitate cosmic ray removal .
image reduction was carried out using iraf and aips .
the individual images were bias - subtracted and flat - fielded using twilight exposures .
the images were then registered using shifts determined from the peak positions of several stars on the ccd near the object and coadded after cosmic ray removal . to improve the signal to noise , the resulting image was smoothed with a gaussian function having a full width half maximum of @xmath24 ( 3 pixels ) giving a final resolution , as measured from stars on the final image , of @xmath7 .
several stars that were visible on both the @xmath1-band image and the narrow band image were used to register the two images to an accuracy of @xmath37 .
absolute astrometric calibration was carried out using the guide star catalogue ( gsc ) image processing system of the space telescope science institute ( @xcite ) resulting in an accuracy of @xmath38 .
infrared spectroscopy was performed on june 14 , 1994 , with the united kingdom infrared telescope ( ukirt ) .
the [ oiii ] @xmath394959,5007 and h@xmath40 @xmath394861 lines , redshifted into the @xmath41 band window , were observed using the cgs4 spectrograph in the long camera mode , giving a pixel scale along the slit of @xmath20 and a spectral resolution of 0.006 micron ( @xmath42 km s@xmath5 ) .
the object was observed in the beam - switching mode : the object was switched every 3 minutes between two different positions on the slit .
the two series of observations were subtracted from each other to produce a sky - subtracted two - dimensional spectrum with a positive and a negative spectrum of the object .
further details of this observing method are described in eales & rawlings ( 1993 ) .
the spectrum was extracted in a @xmath43 aperture ( similar to that used for the optical low resolution spectroscopy ) to include all extended line emission and wavelength and flux - calibrated using the `` longslit '' package as incorporated in iraf .
a @xmath41-band image was obtained and reduced for us by steve eales and steve rawlings using ircam at the united kingdom infrared telescope at mauna kea , hawaii , in march 1993 .
it has a pixel scale of @xmath44x@xmath44 and the total on source integration time was 1750 seconds . for more details on the ircam observations and reductions
see eales & rawlings ( 1995 ) .
the low resolution optical spectrum showed one bright emission line ( fig . 1 ) .
this line is spatially extended by @xmath28 , has an observed equivalent width of 1800 @xmath45300 and an integrated flux of [email protected] x 10@xmath46 erg s@xmath5 @xmath47 in the @xmath28 aperture .
no further emission lines were seen in this spectrum .
the faint continuum of 1243 + 036 is detected redward of the emission line .
the continuum is not detected blueward of the emission line , giving a lower limit to the continuum drop of a factor 2.5 @xmath451 .
the large equivalent width of the line and the drop in continuum level at the emission line ( similar to the continuum drop across the ly@xmath2 emission line in distant quasars due to intervening ly@xmath2 forest absorption , e.g. steidel & sargent , 1987 , and references therein ) along with the faint @xmath1 band identification and the small ultra steep spectrum radio source led us to tentatively identify the line with lyman @xmath2 at @xmath48 .
the redshifted ly@xmath2 line lies close to the strong 5577 skyline making a more exact redshift determination difficult from the low resolution spectrum .
the infrared spectroscopic observation showed two bright emission lines in @xmath41 band ( fig . 2 ) .
these were identified with the [ oiii]@xmath494959,5007 emission lines , indicating that the supposed redshift is indeed correct .
there is also a hint of h@xmath40 @xmath394861 emission .
the [ oiii]5007 has a flux of [email protected] x 10@xmath46 erg s@xmath5 @xmath47 , comparable to that of ly@xmath2 , and has a spatial extent of at least @xmath50 , i.e. comparable to the extent of ly@xmath2 in the low resolution optical spectrum .
the spectrum of the emission lines in the infrared has much lower signal to noise ratio than that of ly@xmath2 in the optical .
therefore any faint [ oiii ] emission in the region outside the radio lobes , if the [ oiii ] would be equally extended as the ly@xmath2 ( see below ) , could not be detected .
figure 3a shows the total intensity image at 8.3 ghz with 0.23@xmath21 resolution .
figure 3b shows the image at 4.7 ghz with a resolution of 0.43@xmath21 and fig .
4 shows an image at 1.5 ghz with @xmath32 resolution , produced from the combined data from rttgering et al .
( 1994 ) , and later observations at the same frequency .
a number of features have been labeled for reference .
table 2 lists the positions of the peaks , the peak surface brightnesses ( i ) and integrated fluxes ( s , measured in a rectangular box around the components ) for the various source components .
the detailed source structure is best represented in the 8.3 ghz image .
the northern lobe ( designated ` a ' ) has a compact component with a ` tail ' extending to the north .
the southern lobe ( designated ` b ' ) again has a compact component ( b1 ) close to the nucleus , and a linear series of knots ( the southern ` jet ' ) , extending towards the south from b1 , with a slight extension to the west at its southern end . in between these two lobes
is a faint region of emission ( designated ` n ' ) , that has a core
jet structure with a length of about 0.8@xmath21 , oriented at a position angle of @xmath51 , i.e. roughly aligned with the axis defined by components a1 , n , and b1 .
the southern jet has an angle of 30@xmath26 away from this axis .
the spectral index image of the 8.3 ghz and 4.7 ghz maps convolved to 0.43@xmath21 resolution ( with a cut - off where the signal to noise ratio in the original maps is smaller than 8) is shown in fig . 5a , and of the 4.7 ghz and 1.46 ghz maps convolved to @xmath32 resolution
is shown in fig .
spectral index values between 8.3 ghz and 4.7 , ghz determined from the surface brightness of the various source components convolved to the same resolution , are also listed in table 2 .
most components of the radio source have steep spectra , with indices typically @xmath52 .
the two exceptions are components b1 and n , both of which have flatter spectral indices of @xmath53 .
one curious aspect of the 1243 + 036 radio structure is that the spectral index distribution steepens with distance from the source centre in both components a and b. this is unlike most high power radio galaxies , which show the flattest spectral indices at the radio hot spots situated at the extremities of the radio source ( @xcite ) .
this spectral steepening may be due to synchrotron radiation losses of the highest energy electrons .
we assume minimum energy conditions ( see miley 1980 ; pacholczyk 1970 ) to derive a minimum magnetic field strength . in calculating the minimum energies we assume that the radio source is cylindrically symmetric
, the plasma has unit filling factor and that the synchrotron spectrum extends from @xmath54 to @xmath55 ghz . for the components with high frequency spectral index
@xmath56 a two slope power - law spectrum was assumed with a spectral index below 4.7 ghz ( observed frame ) of @xmath22 , equal to the integrated source spectral index between 178 mhz and 2700 mhz .
the derived parameters for the various components are in table 3 .
we can estimate the density of the external medium assuming that the minimum pressures are balanced by either ram pressure , due to the propagation of the radio plasma at a velocity of a few thousand km s@xmath5 ( e.g. cygnus a , carilli et al .
1991 and references therein ) , or by thermal pressure from a hot ( @xmath57 k ) external medium , which is often found to be present around powerful radio galaxies and in clusters at low redshifts ( see crawford et al . 1988 ; baum et al .
1989 and references therein ) .
for the typical source parameters , we find that the external density ranges from 0.1 cm@xmath5 ( ram pressure ) to several tens cm@xmath5 ( thermal pressure ) , see table 4 . assuming that the spectral steepening of the outer components relative to the inner components is due to synchrotron radiation losses
, we crudely estimate the synchrotron age of the emitting radio plasma in a few components from the break frequency where the radio spectrum steepens and the magnetic field strength .
we estimate the restframe break frequency by fitting a `` kardashev - pacholczyk model '' to the spectral index distribution in figs . 5a and 5b and assuming that the input spectral index @xmath59 ( spectral index at frequencies below the synchrotron break ) . for more details about spectral ageing modelling
see carilli et al .
the synchrotron age of the radio plasma is given by : @xmath60 myr with b in @xmath61 g and @xmath62 in ghz , see carilli et al .
we note however that there are large uncertainties in these ages , because of the strong dependence on the magnetic fields that are derived from the high restframe frequencies only . in any case , the derived synchrotron ages give an impression of the age of the outer components ( b4 ) relative to the inner components ( b1 ) .
the derived synchrotron ages for the components a1 , b1 and b4 are listed in table 3 .
images of polarized intensity from 1243 + 036 are shown in fig .
6 , along with vectors representing the position angle of the electric field vectors .
the only component that shows substantial polarization at both frequencies is the southern jet ( components b2 , b3 , and b4 ) .
component a shows weak polarization at 8.3 ghz , and no polarized emission at 4.7 ghz , and components b1 and n show no polarized emission at either frequency .
the fractional polarization of these various components ( 4@xmath63 upper limit for b1 ) are listed in table 2 .
also listed are the depolarization ratios , d@xmath64 , for the components with detected polarized signal , where d@xmath64 is the ratio of fractional polarization at 4.7 ghz to that at 8.3 ghz .
component b4 shows no significant depolarization between 4.7 ghz and 8.3 ghz , while component b2 has a value of d@xmath64 = 0.55 , and a1 only has an upper limit ( 4@xmath63 ) of d@xmath64 @xmath65 0.3 .
we estimate that the uncertainties in observed position angle in the brighter polarized regions are about 2@xmath26 , based on signal - to - noise and possible systematic errors of @xmath81@xmath26 discussed in sect . 2 .
overall , we feel that a position angle change of @xmath66 6@xmath67 between 8.3 ghz and 4.7 ghz could have been measured with this data at the peaks in polarized intensity .
no such change is seen , implying an upper limit to the rest frame rotation measures of @xmath81000 rad m@xmath68 .
this limit applies to the regions that are not depolarized . the projected magnetic field distribution is shown in fig . 7 overlayed on the 8.3 ghz total intensity map . in the southern region ,
the projected field is parallel to the jet , and then bends towards the west in the vicinity of knot b4 . in the north
the projected field is parallel to the ` tail ' of component a towards the northwest . towards the west of knot b4
there is slightly extended radio emission visible at all three radio frequencies .
this emission and the change of the magnetic field direction at that position suggest that this may be outflow from the final hotspot ( b4 ) into the radio lobe .
the @xmath1 band image identifying the radio source is from rttgering et al .
( 1995b ) .
the optical identification is a faint @xmath25@xmath69 extended galaxy with an integrated magnitude of @xmath70 ( @xcite ) . in the low resolution optical spectrum , continuum emission is detected but no emission lines were detected in the @xmath1 band region of the spectrum .
the strongest emission line that might affect the @xmath1 band flux is civ@xmath391549 ( expected at 7079 ) but was not detected , so that the @xmath1 band does not contain any appreciable contribution ( less than 1% ) from line emission .
the @xmath1 band image therefore represents the continuum emission from 1243 + 036 .
the optical galaxy is clearly aligned with the main radio axis , as is common in high redshift radio galaxies ( @xcite ) . in fig .
8 the same @xmath1 band image is rebinned and smoothed with a gaussian of fwhm @xmath71 and overlayed with the contours of the 8 ghz vla map
. there appears to be faint extended continuum emission in the direction of the principal radio axis that extends beyond the radio lobes ( indicated a and b in fig .
although at a low surface brightness , the total extended emission is detected at a signal to noise of @xmath810 in both north ( a ) and south ( b ) .
three more small blobs at the 10@xmath63 significance level are seen on the smoothed frame ( indicated c , d and e ) in the region to the southwest of 1243 + 036 , while elsewhere on the ccd frame such a concentration of emission features is not observed . the @xmath41-band image obtained at ukirt gave a magnitude for 1243 + 036 in a @xmath50 aperture of @xmath41=18.8
@xmath45 0.2 .
however , the @xmath41-band flux includes the strong redshifted [ oiii ] emission line ( see fig . 2 ) .
after correcting for the [ oiii ] flux as determined from the infrared spectroscopy , we find a continuum @xmath41 magnitude of 19.3 @xmath45 0.3 .
this @xmath41-band magnitude is typical for radio galaxies at these redshifts ( e.g. eales & rawlings 1993 ; eales et al .
the signal to noise ratio of the detection was low and the image is therefore not shown .
a keck - telescope image in @xmath41 band shows that it is aligned with the optical and radio structure ( van breugel , private communication ) . a contour image of the lyman @xmath2 emission from 1243 + 036 is shown in fig .
the peak , measured as the position of the maximum of a twodimensional parabolic fit to the brightness distribution , is at 12@xmath72 45@xmath73 38.36@xmath74 , 03@xmath67 23@xmath75 21.1@xmath21 .
the ly@xmath2 emission has a complex morphology : ( 1 ) ` cone - shaped ' emission extends along the radio axis towards the south
east for about 2@xmath21 .
( 2 ) on a scale of @xmath76 the gas shows two ` arms ' of emission both extending almost due south , originating from the southern end of the inner cone of emission . ( 3 ) to the north there is faint emission which extends about 3@xmath21 from the ly@xmath2 peak at a position angle of about @xmath77 . ( 4 ) the lowest surface brightness ly@xmath2 emission on scales of @xmath78 has a clumpy or filamentary structure .
this clumpy structure is at a level of 46@xmath63 .
( 5 ) there is a small region of enhanced line emission at the southern end of the ly@xmath2 emission , due south of the ` gap ' between the two ` arms ' , mentioned above . the positions of the slits used in the 2.8 resolution spectroscopy are shown in fig .
the spectrum with the slit along the radio axis ( pa1=152@xmath26 ) shows the velocity field of the central and most spatially extended ly@xmath2 .
the overall shape of the central velocity profile ( fig .
10 ) has a width of @xmath79 km s@xmath5 ( fwhm ) .
the line does however not have a smooth shape .
it has several small dips ( or peaks ) , especially in the blue wing of the ly@xmath2 emission line .
these small dips at 5533 , 5538 , 5544 and 5550 were present in the individual exposures . the two - dimensional ly@xmath2 spectrum in pa1 , smoothed with a gaussian of fwhm 2 pixels ( @xmath80 ) to enhance fainter extended ly@xmath2 emission , is shown in fig . 11 .
most of the flux is in the centre and has a spatial extent of @xmath81 that peaks at 5557 i.e. @xmath82 .
the total flux of ly@xmath2 is better determined from the high resolution spectrum , since the subtraction of the skyline at 5577 can be carried out with a much greater accuracy than in the low resolution spectrum .
the integrated ly@xmath2 flux in an aperture of @xmath83 is @xmath84 x 10@xmath46 erg s@xmath5 @xmath47 corresponding to a luminosity of @xmath85 x 10@xmath86 erg s@xmath5 . in a larger aperture of @xmath87 , including the most extended emission seen in the smoothed two - dimensional spectrum ( see below and fig .
11 ) , the ly@xmath2 flux is @xmath88 x 10@xmath46 erg s@xmath5 @xmath47 .
a separate velocity component in ly@xmath2 emission is seen at about 1100 km s@xmath5 blueward and @xmath25 southward along the slit from the central ly@xmath2 peak .
this component contains almost 10% of the total ly@xmath2 flux and has a velocity width of @xmath89100 km s@xmath5 ( fwhm ) .
its position coincides with the brighter of the two ` arms ' of ly@xmath2 emission seen in the ly@xmath2 narrow band image and with radio component b1 ( see below ) .
there is also a remarkable , extended narrow component ( @xmath90 km s@xmath5 fwhm ) that clearly shows a velocity shear on both sides of the nucleus .
it has a total spatial extent of @xmath91 , corresponding to 136 kpc , much larger than the size of the radio source ( @xmath78 ) .
the velocity difference between the northern and southern extent of this emission is @xmath92 km s@xmath5 .
note that this very extended ly@xmath2 component was not seen in the narrow band ly@xmath2 image .
this is because the resolution of the spectrum is very well matched to the velocity width of the most extended ly@xmath2 gas , whereas the narrow band filter has a much larger spectral width .
also the presence of the 5577 skyline adds a considerable amount of noise in the narrow band image .
the small dips seen in the one - dimensional spectrum are also clearly visible on the two - dimensional spectrum ( fig .
11 ) . especially noteworthy
is the feature at @xmath93 , that extends spatially from the central ly@xmath2 emission to the separate velocity component at b1 . the high resolution spectrum taken with the slit perpendicular to the radio axis ( pa2=62@xmath26 , figs .
12 and 13 ) shows only the ly@xmath2 emission from the brightest inner part as seen on the ly@xmath2 narrow band image . the emission is spatially resolved .
the emission has the same overall velocity width and peaks at the same wavelength as the spectrum taken along the radio axis .
the two - dimensional spectrum shows a depression in the ly@xmath2 emission blueward of the peak in the central rows of the spectrum over about 6 ( @xmath94 km s@xmath5 ) .
the ly@xmath2 emission blueward of the depression has the same velocity as that of the separate velocity component described above along pa1 .
redward of the depression the ly@xmath2 has the same velocity as that of the brightest inner ly@xmath2 in the pa1 spectrum .
the signal of the main ly@xmath2 emission is not as strong as in the pa1 spectrum .
therefore it seems that the @xmath31 wide slit was not exactly positioned at the peak of ly@xmath2 and the blueshifted ly@xmath2 emission at the position of b1 is also at least partly included ( see fig .
the very extended low surface brightness ly@xmath2 seen in the spectrum along the radio axis ( pa1 ) is not seen in the spectrum with the perpendicular slit position ( pa2 ) . although the spectrum at pa2 is not as deep as that at pa1 ( the noise is a factor 2 higher ) , such faint emission would have been detected if it was of the same extent and surface brightness as the emission as observed along the radio axis .
figure 14 shows the lyman @xmath2 image of 1243 + 036 overlayed with the radio continuum emission at 8.3 ghz .
figure 8 shows the same radio image overlayed with the @xmath1 band optical image .
standard vla astrometry is accurate to at least 0.2@xmath21 , the optical astrometry has an accuracy of @xmath95 ( @xcite ) .
the spatial extent of the ly@xmath2 emission visible in the narrow band image is similar to that of the radio source ( @xmath88@xmath21 ) .
component n of the radio source is located at the peak of the line emission and @xmath1-band emission . since n is also the flattest spectrum radio component ,
we identify it as the centre of the nuclear activity of 1243 + 036 .
the spatial extent of this radio core jet component ( length @xmath96 0.8@xmath21 ) is comparable to the spatial extent of the highest surface brightness line emission . the radio galaxy 1243 + 036 exhibits two trends seen in samples of high redshift radio galaxies ( _ cf .
_ mccarthy 1993a ) .
firstly , the alignment between the major axis of the line emitting gas and the optical continuum with the principal axis of the radio source . secondly , the emission line gas is distributed asymmetrically with respect to the nucleus . in most powerful radio galaxies
the emission line gas is brightest on the side of the closest radio lobe , suggesting interaction between the radio source and the ambient gaseous medium ( @xcite ) .
regarding the entire extent of the radio lobes of 1243 + 036 , the southern radio lobe ( b ) is not the closest to the nucleus ( n ) .
however , regarding the brightest hotspots in the lobes , a1 and b1 , b1 is the closer with an armlength ratio of q=1.2 further , the southern lobe is the most distorted and brightest , indicating a stronger interaction with the environment .
in addition to these two trends , there are three further correlations in the radio and ly@xmath2 properties .
firstly , both the line emitting gas and the radio source show peculiar structure at the position of component b1 . the radio source bends by about 30@xmath26 towards the south .
the line gas has a local peak at this position , and the long slit spectra show that the line emitting gas in this region is blueshifted by @xmath97 km s@xmath5 with respect to the central ionized gas .
the southern jet follows closely along the brighter of the two southern ` arms ' of line emission .
secondly , the emission line gas is seen to encompass the northern radio source component a. the high resolution spectrum ( fig . 11 ) shows that the ly@xmath2 has a low surface brightness component that extends far beyond the radio lobes .
thirdly , the emission line gas inside the radio structure is brighter and has a much larger velocity dispersion than the emission line gas outside the radio structure . summarizing the results from the high resolution spectroscopy and narrow band imaging of the ly@xmath2 gas we can distinguish between three separate components of the emission line gas : \(i ) the high velocity dispersion gas inside the radio structure , the _
`` inner halo''_. the ly@xmath2 emission within a radius defined by the maximum extent of the radio structure is dominated by the bright central emission which has a shape of a cone with a large velocity dispersion ( @xmath98 km s@xmath5 fwhm ) .
such a velocity dispersion is typical for the extended emission line gas in high redshift radio galaxies ( @xcite ) .
\(ii ) enhanced ly@xmath2 emission with a blueshifted velocity at the position of radio knot b1 .
the region of enhanced ly@xmath2 emission at the position of radio knot b1 , in the brighter of the two southern arms , has a clear velocity difference ( @xmath97 km s@xmath5 ) relative to the central ly@xmath2 .
its velocity dispersion ( @xmath99 km s@xmath5 fwhm ) is similar to the rest of the ly@xmath2 gas in the inner halo .
\(iii ) ly@xmath2 emission extended far beyond the radio source , the _ `` outer halo''_. the very extended ly@xmath2 emission with a velocity shear of @xmath100 km s@xmath5 has a total spatial extent of @xmath6 ( 136 kpc ) , much larger than the radio source ( @xmath101 ) .
this indicates that this component of the gas has no direct relation with the radio plasma .
this extended ly@xmath2 was seen with the slit of the spectrograph along the principal radio axis .
the spectrum perpendicular to the radio axis showed that in this direction there is no ly@xmath2 emission of such extent .
several mechanisms have been considered for the ionization of the extended emission line clouds in powerful radio galaxies . these include ionization by uv radiation from hot stars , shock ionization from the interaction with the radio lobes and photoionization by the active nucleus .
the equivalent width of the ly@xmath2 emission in most high redshift radio galaxies is several times larger than can be produced by hot stars ( @xcite ) . in general , the line ratios in radio galaxies are not well reproduced by shock ionization nor ionization by hot stars and indicate that photoionization by nuclear uv radiation appears to be the dominant mechanism of ionization ( @xcite ) .
anisotropic photoionization by an obscured nucleus has been argued to explain the deficit of observed uv photons from radio galaxies and seyfert-2 galaxies relative to the amount needed to produce the emission line luminosities ( @xcite ) .
this concept has been an important constituent of orientation - based unification models of radio galaxies and quasars were developed , which can explain many observed correlations and anomalies ( @xcite ) . in these models
radio - loud quasars and radio galaxies are intrinsically similar objects but observed at different angles .
the radiation from the nucleus escapes preferentially within a cone along the radio axis and ionizes the gas clouds in its path . in quasars
we can see inside the ionizing cone and see the broad - line emission region close to the active nucleus , while in radio galaxies the nucleus is obscured so that we can not see the ionizing radiation directly , but see the extended ( narrow ) emission line region that is ionized by the nuclear continuum .
the alignment of the ionized gas with the radio axis in most powerful radio galaxies argues in favour of these models , especially the presence of ionized gas outside the radio lobes .
the ly@xmath2 emission in the outer halo in 1243 + 036 , outside the radio lobes , but aligned with the principal radio axis , rules out shock ionization by the radio source and strongly argues for photoionization by anisotropic nuclear emission .
4c41.17 at z=3.8 ( @xcite ) and 6c1232 + 39 at z=3.2 ( @xcite ) are other examples of high redshift radio galaxies with ly@xmath2 emission extending beyond the extent of the radio emission .
the `` cone '' shaped morphology of the innermost @xmath25 of the ly@xmath2 emission of 1243 + 036 would not be expected from shock ionization and further supports the idea that the ly@xmath2 halo of 1243 + 036 is indeed photoionized by anisotropic nuclear continuum radiation .
the shape of this innermost ly@xmath2 emission region indicates an opening angle for the cone of nuclear ionizing radiation of between 35@xmath26 and 50@xmath26 .
although we have very little information on the physical conditions of the ionized gas we can estimate the density and mass of the emission line gas , assuming photoionization , from the flux and extent of the ly@xmath2 emission along the lines of mccarthy et al .
( 1990 ) for 3c294 and chambers et al .
( 1990 ) for 4c41.17 .
we shall use three different approaches to estimate the density .
i. following mccarthy ( 1990 ) , the ly@xmath2 luminosity is related to the density of the ionized gas , assuming case b recombination : @xmath102x@xmath103 ergs s@xmath5 , where @xmath104 is the total volume occupied by the emission line gas and @xmath105 the volume filling factor of the gas . from this
then follows the mass in ionized gas m@xmath106 .
because we have no means of measuring the filling factor and density simultaneously in 1243 + 036 , we have to estimate the volume filling factor . from direct measurements , using sulphur lines , of the density of line emitting gas in low redshift radio galaxies , filling factors in the range 10@xmath10710@xmath108 and typically of order @xmath109 have been found ( @xcite ) . assuming a filling factor @xmath110 we give in table 5 the density of several components of the ly@xmath2 halo and their inferred masses .
the sizes of components were measured from the ly@xmath2 image and high resolution spectroscopy and were assumed to be cylindrically symmetric .
the values derived in table 5 indicate that the density of the ionized gas falls off with radius as @xmath111 .
this would mean that in the case of anisotropic nuclear photoionization the ionization parameter would be about the same at all radii .
this is consistent with the findings of baum et al .
( 1992 ) for the emission line nebulae in powerful radio galaxies at low redshifts and also found for the extended line emission around quasars at @xmath112 ( see bremer et al .
1992a and references therein ) .
the ly@xmath2 peak at b1 seems to be an exception .
however , as we will argue in sect .
4.3 , the jet gas interaction that bends the radio jet at the position b1 , may have shredded dense neutral cores of the emission line clouds , therefore increasing the amount of gas exposed to the nuclear ionizing radiation and increasing the volume filling factor of ionized gas , consequently enhancing the ly@xmath2 emission at that position ( see bremer et al .
if the volume filling factor is locally increased to 10@xmath113 , the density decreases to @xmath8100 cm@xmath113 , similar to that in the rest of the inner halo .
we note that this would infer a higher mass ( @xmath115 m@xmath116 ) of ionized gas at b1 .
\ii . to prevent the emission line clouds from dispersing on short time scales ,
they must be confined by external pressure , as was pointed out by fabian et al .
we can estimate the density of the emission line filaments , assuming that they are confined by a hot ( @xmath117 k ) external medium , with density derived from the radio data .
the emission line clouds may also be confined by the pressure from the radio lobes , however the radio lobes in turn must be confined by an external medium , leading to the same pressure requirement .
further evidence for the presence of a hot halo comes from the strong depolarization of the radio emission , that is most likely caused by a hot ( x - ray ) halo as magnetoionic medium ( @xcite ) .
other hzrgs ( e.g. 4c41.17 and b20902 + 34 , carilli et al .
1994 ) show evidence for the presence of a hot surrounding halo .
unfortunately , at the high restframe frequencies observed , the radio emission from the extended lobes is not detected and the emission is dominated by the emission from the hotspots .
the hotspots are most likely confined by the ram pressure caused by their propagation through the external medium and not by static thermal pressure . because of its flatter radio spectrum hot spot b1 appears to be the most `` active '' and we therefore assume the balance between minimum pressure and ram pressure from b1 to be the best representative pressure
. ram pressure confinement of the advancing hotspot requires a density of @xmath80.1 cm@xmath113 ( see table 4 and sect .
hotspot a1 requires a similar density . assuming pressure balance between the hot halo gas ( @xmath118 k cm@xmath113 ) and the emission line clouds ( with a temperature of @xmath119 k ) gives an average density of the ionized gas of @xmath8100 cm@xmath113 , roughly consistent with the emission line gas density derived assuming @xmath110 .
\iii . assuming that every ionized hydrogen atom eventually results in the emission of a ly@xmath2 photon , the ly@xmath2 luminosity indicates the necessary flux of ionizing photons , q@xmath120x10@xmath121 photons s@xmath5 .
using this and an estimate of the ionization parameter , u , of the gas gives a third estimate of the density : u = q@xmath122 . in 1243 + 036
we have no emission lines from which we could directly estimate the ionization parameter .
however , from photoionization modelling of high and low redshift radio galaxies , the best fit to the emission line ratios is achieved for log u=@xmath123 ( @xcite ) .
a similar number is obtained in 3c quasars where q is known ( @xcite ) .
thus , assuming log u=@xmath123 at a characteristic radius in the inner halo of @xmath825 kpc gives a density @xmath124 3 cm@xmath113 .
however this assumes the ionizing radiation is emitted isotropically .
if photoionization is anisotropic within a cone of opening angle @xmath845@xmath26 , the ionizing flux and corresponding densities are an order of magnitude higher .
we note that the derived densities are only crude estimates due to the many uncertainties ( e.g. filling factor and minimum pressures ) and can therefore be only order of magnitude estimates .
we shall discuss the kinematics of the emission line gas for each of the components as defined in sect .
3.5 separately .
\(i ) _ inner halo _ if the gas motion of the inner halo is gravitational in origin , the velocity dispersion would imply a virial mass of @xmath125 m@xmath116 inside a radius of @xmath830 kpc .
however , the spatial correlation of the high velocity gas of the inner halo and the radio structure , the low velocity dispersion of the outer halo and the enhanced and accelerated ly@xmath2 gas at radio knot b1 , suggest that the high velocity dispersion of the inner halo gas is the result of hydrodynamical interactions with the radio plasma .
the interaction of emission line gas with the radio source has been suggested earlier as the cause for the high velocity dispersions of the extended emission line gas in high and intermediate redshift radio galaxies and radio loud quasars and there is evidence that radio sources can entrain and accelerate gas ( @xcite ) .
calculations by de young ( 1986 ) showed that gas entrainment rate can be up to a few hundred solar masses per year .
the narrow dips , observed mainly on the blue wing of the ly@xmath2 profile of the inner halo and at least one extending spatially from the central part to the separate velocity component at b1 , are interesting features that deserve some attention .
although we can not exclude that these features are due to true velocity structure of the line emitting gas , the spatial continuity of one of these dips from the central ly@xmath2 to the blueshifted ly@xmath2 emission component at b1 indicates that they are probably due to absorption by associated neutral hydrogen clouds .
this also indicates that these absorption clouds must be in front of the ly@xmath2 emission region and have a minimal spatial extent of @xmath126 kpc .
if the emission line gas is ionized by anisotropic nuclear photoionization as in the unification models of quasars and radio galaxies , the absorption systems seen against the ly@xmath2 emission in 1243 + 036 may be clouds similar to the ly@xmath2 emitting clouds , situated outside the cone of ionizing radiation , but in our line of sight to the ly@xmath2 gas ( @xcite ) .
\(ii ) _ enhanced ly@xmath2 at b1 _ the positional coincidence of the separate velocity component of the emission line gas and the bend in the radio structure at b1 suggests a direct interaction of the radio jet and the ly@xmath2 gas accelerating the gas to its obserrved velocity .
the interaction of the radio jet with the gas could have locally enhanced the ly@xmath2 emission .
this will be further discussed in sect .
\(iii ) _ outer halo _ the narrow velocity width ( 250 km s@xmath5 fwhm ) of the emission line gas in the outer halo compared to that of the ly@xmath2 emission in the inner halo and its extent to far beyond the radio source supports the suggestion that this very extended component is not affected by the interactions that have stirred up the inner ly@xmath2 gas , as mentioned above .
the velocity shear observed in the outer halo of @xmath92 km s@xmath5 may be due to rotation or a flow of gas , i.e. infall or outflow .
we will discuss these possibilities and their possible connections and implications for the formation of powerful radio galaxies in the early universe .
_ a ) rotation _ the velocity shear might be due to a large scale rotation of a gas disk in the halo .
a gravitational origin of the rotation of such a large disk implies a mass of @xmath125 sin@xmath127 m@xmath116 , where @xmath128 is the inclination angle of the disk with respect to the plane of the sky .
baum et al .
( 1992 ) propose that at low redshifts rotation of the extended emission line gas in powerful radio galaxies indicates the presence of dynamically young disks of gas acquired in a recent interaction or merger with a gas - rich galaxy .
these rotating emission line regions are , however , only on scales of @xmath820 kpc , while in 1243 + 036 the outer halo is almost an order of magnitude larger .
large gaseous halos are common around hzrgs ( @xcite ) and may imply that all primeval galaxies in the direct ( cluster ) environment of hzrgs also possessed such large halos .
an interaction between two ( primeval ) galaxies with large gaseous halos may leave kinematical signs of the event over a much larger scale than the relatively small rotating emission line nebulae that exist around radio galaxies at low redshifts that according to baum et al .
( 1992 ) are due to interactions .
thus , it is possible that the rotating outer halo of 1243 + 036 originates from an interaction with a neighbouring primeval galaxy . alternatively and
perhaps more likely , the outer halo ly@xmath2 emitting gas in 1243 + 036 may be a relic of the gaseous halo from which the galaxy has been forming since before the radio source switched on .
a rotating disk may have formed during the accretion of gas from the environment .
numerical simulations by evrard et al .
( 1994 ) of the formation of galaxies in hierarchical clustering scenarios indicate that rotating disks with radii of several tens of kiloparsecs naturally form during the formation of galaxies at high redshift ( @xmath1291 ) .
note that , although the gas disk does not have to be rotating exactly along the radio axis , it must be within the cone of ionizing radiation from the agn ( if it is anisotropically ionized ) and , more importantly , it must be within the slit of the spectrograph that was oriented along the radio axis .
the slit has a width of 17 kpc at the redshift of 1243 + 036 . assuming that a rotating gaseous disk would not be wider than 35 kpc ( one quarter of its observed diameter ) at a radius of @xmath130 from the nucleus , this limits the difference between the orientation of the disk and the radio axis to less than 20@xmath26 .
the fact that we do not see evidence for a disk in absorption against the bright central ly@xmath2 emission does not argue against the explanation of the velocity shear as originating from a rotating gas disk .
if the observed ly@xmath2 gas in the outer halo is ionized by anisotropic nuclear radiation , we would only expect to see the rotating gas disk in absorption against the bright central ly@xmath2 if we saw the disk edge on .
the size ( in the direction of pa2 ) of the bright central ly@xmath2 emission region ( @xmath131 kpc ) implies an angle between the orientation of the disk and our line of sight of at least 4@xmath26 .
it may seem surprising that in the case of 1243 + 036 the rotation axis of the gas disk is perpendicular to the radio source axis , because it is generally believed that radio galaxies have a central black hole fuelled by an accretion disk in the inner parts of the galaxy that has its rotation axis along the black hole s spin axis and the radio axis .
thus , the the rotation of the outer halo in 1243 + 036 can not be a direct extension of the inner accretion disk .
however , west ( 1994 ) notes that there are sometimes remarkable alignments between radio axes of relatively close pairs of powerful radio galaxies , and also between the major axes of neighbouring clusters and their central galaxies .
he suggests that this may be a signature of structure formation in the universe from the largest scale structures down to the scales of the most massive galaxies in clusters . in this case
the preferred axis for structure formation is the line between neighbouring systems .
rttgering et al .
( 1996 ) find an excess of galaxies along the radio axis of powerful radio galaxies .
the orientation of the outer ly@xmath2 halo in 1243 + 036 along the radio axis , would then fit into this scenario of a connection between the ( radio ) axes of massive galaxies and large scale structure formation . _
b ) outflow _ in the case of outflow , it is unlikely that the gas would advance through the extragalactic medium faster than the radio lobes .
therefore , the outflow of gas would need to have existed long before the observed radio source switched on . with the maximum velocity of the most extended emission ( @xmath90 km s@xmath5 ) it would take @xmath132x@xmath133 years to transport the gas out to a radius of 70 kpc , an order of magnitude longer than the most plausible age of the radio source ( @xmath57 years , based on the projected radio size and a radio lobe propagation speed of @xmath134 km s@xmath5 , see carilli et al .
1991 and references therein ) .
however , we may simply see the galaxy during its most recent period of activity and it may have developed an extended radio source several times during previous active periods in the time needed for the outflow to build up the outer ly@xmath2 halo . in that case
the outer halo may be gas swept up and deposited by expanding radio lobes during earlier active periods of the galaxy .
possibly , 1243 + 036 may have had an active nucleus before its radio activity , ejecting gas along the central black hole s rotation axis .
there is evidence for bulk mass outflow of ionized gas close to the nucleus of quasars at velocities up to 1000 km s@xmath5 ( @xcite ) . in broad absorption line
qsos even much higher outflow velocities are observed ( e.g. turnshek et al .
1988 ; turnshek 1988 ) .
however , there have not previously been indications that such outflows originating from the agn might continue at lower velocities out to distances of 70 kpc .
therefore , we regard it as unlikely that gas ejection by the agn is the origin of the outer halo of 1243 + 036 and its velocity shear .
an alternative outflow mechanism may be an equivalent of the ` superwinds ' observed in low redshift strong iras galaxies ( e.g. heckman et al .
1990 ; heckman et al . 1993 ) .
these outflows are believed to be caused by a massive starburst in the galaxy . in the burst of star formation
, the kinetic energy supplied by stellar winds and supernovae results in the expulsion of interstellar material into the surrounding medium , and can drive a large scale outflow . some ultraluminous iras galaxies
( e.g. arp220 , mrk266 ) have superwinds producing emission line nebulae several tens of kiloparsecs in size ( see heckman et al . 1993 and references therein ) .
linewidths and observed outflow velocities of emission line gas in the superwinds are typically a few hundred kilometers per second , similar to what we find for the velocity shear and width of the ly@xmath2 emission in the outer halo of 1243 + 036 .
if a superwind is responsible for the presence and kinematics of the outer halo ly@xmath2 gas in 1243 + 036 , then the timescale of the outflow would mean that a massive initial starburst must have occurred long before the radio source switched on .
heckman et al .
( 1990 ) showed that an outflow velocity @xmath135 km s@xmath5 at a radius @xmath136 kiloparsec with a density of the ambient medium @xmath137 requires a constant energy injection rate @xmath138 . to produce this energy injection rate by a starburst
requires a star formation rate sfr@xmath139 ergs s@xmath5 ) m@xmath116 yr@xmath5 .
the velocity shear and extent of the outer halo in 1243 + 036 implies an energy injection rate of @xmath140 ergs s@xmath5 during a period of the halo s dynamical timescale of @xmath141 years .
if this energy injection rate is indeed produced by a starburst , it implies a star formation rate for 1243 + 036 of @xmath142 m@xmath116 yr@xmath5 assuming a salpeter imf ( heckman et al .
1990 and references therein ) .
a starburst that is biased towards more massive stars may still require 1000 m@xmath116 yr@xmath5 , an order of magnitude more than derived for arp220 ( @xcite ) . over the dynamical timescale of the outer halo , the entire stellar population of a massive galaxy
would have formed .
although such star formations rates are enormous , they are also inferred from stellar population modelling of the colours of high redshift radio galaxies and are invoked to explain the alignment effect of the optical and radio axis of high redshift radio galaxies by jet - induced star formation ( e.g. 4c41.17 , chambers et al . 1990 ) .
such a large population of young hot stars could account for the photoionization of a large part of the ly@xmath2 gas .
_ c ) infall _ the ly@xmath2 gas of the outer halo might be infalling gas from the outer parts of the protogalactic halo . if the radio source associated with 1243 + 036 is oriented at some angle to the plane of the sky , gas infalling along the radio axis and anisotropically ionized by the agn could show a velocity structure similar to that observed for the outer halo ly@xmath2 gas .
however , the radio polarization data of 1243 + 036 show that the southern radio structure is the least depolarized . if 1243 + 036 is embedded in a magnetoionic halo , as appears to be the case for most radio galaxies and quasars ( laing garrington effect , garrington et al .
1988 ; laing 1988a ; carilli et al . 1994 ) , the southern radio jet should be on the side closest to us .
the radiation from the southern part of the radio source has then traveled a shorter distance through the magnetoionic halo than the the northern radio emission and is therefore the least depolarized .
this orientation is further supported by the blueshifted ly@xmath2 component at the position of the southern radio knot b1 , probably caused by the interaction of the radio jet and the gas ( see sect .
therefore , it seems somewhat surprising that ionized gas falling in from large radii on the southern side is also blueshifted with respect to the systemic ly@xmath2 and not redshifted .
thus , the most likely orientation of the radio axis with respect to the plane of the sky argues against the outer halo ly@xmath2 emission being due to infall along the radio axis .
nevertheless , we will consider whether an inflow mechanism could produce the outer halo in 1243 + 036 .
there is much evidence that powerful radio galaxies at high redshifts reside in clusters ( e.g. hill & lilly 1991 and references therein ) , and that they have a hot ( @xmath810@xmath143 k ) `` cooling flow '' intra cluster medium ( e.g. cygnus a , arnaud et al .
1984 ; 3c295 , henry & henriksen 1986 ; 3c356 , crawford & fabian 1993 ; see also fabian 1994 ; carilli et al .
1994 ) . our determination of the minimum pressures and polarization measurements of the radio structure of 1243 + 036 support this idea ( see sect .
if the cooling time of that gas is short compared to the age of the universe at the observed epoch then there will be a cooling flow , depositing cold gas near the centre of the halo . at low redshifts
the cooling time of the hot icm in many clusters is shorter than the hubble time and therefore gas is cooling and deposited , possibly forming emission line clouds . in nearby
cooling flows mass deposition rates up to several hundred m@xmath116 yr@xmath5 have been deduced ( see fabian 1994 and references therein ) .
the mass of the _ ionized gas _ in the outer halo of 1243 + 036 in the observed area that was covered by the slit in orientation pa1 , is of order 10@xmath144 m@xmath116 ( see sect .
4.2 ) . in the case of a cooling flow
, the gas has no preferential direction of infall , so that it is not only along the radio axis ( slit orientation pa1 ) , in the ionization cone from the agn , but roughly isotropically distributed around the radio galaxy .
thus , the total mass of cold ( @xmath145 k ) gas in the outer halo all around the galaxy is minimally of order 10@xmath146 m@xmath116 , situated at a radius of @xmath860 kpc .
if the ionized gas represents only a fraction of the total mass , the total amount of gas in the outer halo may be even an order of magnitude higher . at the redshift of 1243 + 036 the cooling time
would have to be shorter than 10@xmath146 years ( the approximate age of the universe at that epoch ) to allow a cooling flow to be the origin of the emission line gas . from an estimated density of 0.1 cm@xmath113 at a radius @xmath810 kpc ( see sect .
4.1 ) , temperature @xmath147 k and the cooling function from thomas ( 1988 ) , the cooling time is @xmath810@xmath144 years . the density in a hot halo decreases with radius at a rate between @xmath148 and @xmath111 , and the cooling time increases as @xmath149 with constant temperature ( @xcite ) .
then at @xmath150 kpc the cooling time is longer than in the inner part of the ly@xmath2 halo , but is still @xmath810@xmath146 years .
thus it may be possible that a massive cooling flow has deposited a large amount of gas in the outer halo of 1243 + 036 .
a problem for the cooling flow hypothesis is the large velocity shear of the outer halo . in a cooling flow , this is expected to be only a few tens of kilometers per second ( fabian 1994 ; fabian et al .
1987 ) , an order of magnitude smaller than that observed in 1243 + 036 . a flow velocity of only 20 km s@xmath5 means a mass deposition rate of @xmath81000 m@xmath116 yr@xmath5 within a radius of 60 kpc , assuming a density of 0.1 cm@xmath113 .
fabian et al .
( 1987 ) suggests that galaxy interactions are most likely to cause large velocities and keplerian motion of the gas deposited by a cooling flow .
thus even though a cooling flow may be at least in part responsible for the presence of the emission line gas in the outer halo of 1243 + 036 , it can not explain the observed velocity shear .
an additional mechanism , e.g. a galaxy interaction is needed to account for that . however , unless our arguments about the orientation of the radio source are wrong , i.e. the enhanced , blueshifted ly@xmath2 at the projected position of bending point b1 is coincidental and the polarization asymmetry is due to some extraordinary geometry of the magnetoionic halo , we regard the infall of gas as the most unlikely explanation of the velocity shear . in summary , we think that rotation due to accretion of gas associated with the formation of the galaxy is the most plausible origin for the presence and kinematics of the outer halo , but outflow in the form of a ` superwind ' may also be feasible
. a cooling flow may contribute to the presence of the gas but can not be the cause of the velocity shear .
infall in general seems excluded by the most likely orientation of the radio source .
regardless of the actual origin of the gas in the outer halo and its velocity shear , it is clear that it must have been there since before ( the currently active period of ) the radio source switched on .
baum et al .
( 1992 ) classified low redshift radio galaxies on the basis of the kinematics of the emission line gas into `` rotators '' , which baum et al .
( 1992 ) associate with mergers , `` calm non - rotators '' , identified with cooling flows , and `` violent non - rotators '' , where any organized motion of the emission line gas is overshadowed by interaction with the agn .
1243 + 036 shows the presence of two of the kinematical characteristics of the low redshift radio galaxies : a calm rotator in the region outside the radio source and a violent non - rotator in the inner parts .
thus , we may in 1243 + 036 be witnessing an important stage in the evolution of massive galaxies in the early universe . the scenario would be that a large scale rotating gas disk or outflowing gas is ionized by the strong continuum emission from the agn , while the interaction of the radio source with the gas causes large velocities removing signs of the earlier calm kinematics as the radio lobes propagate outwards . in the outer halo , where the gas is still unaffected by the radio plasma ,
the rotating gas disk is still intact .
we note that in other high redshift radio galaxies a strong extended hi absorption feature against the ly@xmath2 is frequently observed ( @xcite ) .
these absorption features have velocity widths of a few hundred kilometers per second and absorb a part of the emission line profile over the entire ly@xmath2 spatial extent along the radio axis .
they may be caused by high column density ( @xmath151 cm @xmath68 ) hi absorption systems with a doppler parameter @xmath152 km s@xmath5 ) or by a superposition of many lower column density clouds with an integrated velocity dispersion like that observed of these absorption features .
if the latter explanation is the case , the similar velocity width of the strong absorption features observed in some hzrgs and of the ly@xmath2 emission from the outer halo in 1243 + 036 , suggests that these features may have a similar origin .
finally we note that the low surface brightness of the ly@xmath2 emission from the outer halo of 1243 + 036 is of a similar level to that occasionally observed from the extended nebulosity around radio quiet quasars ( e.g. bremer et al .
1992b ) and from some damped ly@xmath2 absorption systems ( e.g. pettini et al .
perhaps this represents the quiescent state of the gas around primeval galaxies in general , unaffected by interaction with radio plasma .
several authors have discussed multiple hotspots and bending of the radio jet in quasars and radio galaxies ( e.g. van breugel et al .
1985 ; lonsdale & barthel 1986a , 1986b ; cox et al . 1991 ; icke 1991 )
. an important feature of the bending of the radio jet in 1243 + 036 is the coincidence of enhanced ly@xmath2 emission that is at a blueshifted velocity with respect to the rest of ly@xmath2 emitting gas , suggesting a direct interaction between the radio jet and the emission line gas .
line emission at the position of a sudden bend in the radio jet is also seen at low redshift in 3c277.3 ( coma a ) ( @xcite ) .
we will consider three possibilities for the bent radio structure in 1243 + 036 .
\(i ) _ jet deflection by a gas cloud . _
lonsdale & barthel ( 1986 ) argue that the double hotspots often observed in powerful radio sources are due to deflection of the radio jets by a gas cloud .
they find that a straight deflection by a dense wall of a massive cloud would be possible , but not very appealing , because very large cloud masses are required ( @xmath15310@xmath146 m@xmath116 ) .
also , the jet would quickly eat into the wall so that the deflection direction would change more rapidly than the necessary time to build up the secondary hot spots .
lonsdale & barthel ( 1986 ) find that the best possible method of deflection might be through a `` delaval '' nozzle ( e.g. blandford & rees 1974 ) .
the jet eats into a cloud , where it inflates a plasma bubble .
the bubble breaks out at the weakest spot in the asymmetrically confining cloud , creating the `` delaval '' nozzle .
this then produces collimated outflow that feeds the secondary hotspot .
they claim that the cloud with bubble structure will have a long lifetime such that the deflection direction is maintained stationary over the time required to build up the secondary hotspots ( @xmath810@xmath154 years ) .
the mass and density of the deflecting cloud must be sufficiently large such that it is not blown away past the secondary hot spot . from the total luminosity of 1243 + 036 , the average energy supply rate to the hot spots is unlikely to be larger than @xmath155 ergs s@xmath5 ( see table 3 ) . from this and the total minimum energy of component b4
, we estimate that the time to build up component b4 is 10@xmath15610@xmath154 years ( see also lonsdale and barthel , 1986a ) .
thus , the advance speed ( @xmath157 ) of hotspot b1 must be @xmath158 so that b1 does not pass the other hotspots within that time . from minimum energy density ( @xmath159 ) for b1 in the ram - pressure balance equation @xmath160
, the density of the deflecting cloud must be @xmath161 cm@xmath113 ( see also table 4 ) , similar to what lonsdale and barthel ( 1986 ) derived .
this @xmath162 cm@xmath113 density gas represents the hot halo as suggested earlier and is sufficient to provide the necessary time to build up the secondary hotspots . as this hot gas
most likely uniformly surrounds the whole system it does not itself deflect the radio jet .
it must be an overdensity of hot and cold gas , e.g. emission line gas , that deflects the radio jet .
this whole region would then be accelerated by the interaction and we shall refer to it as the deflecting `` cloud '' . the deflecting cloud must be larger than the size of the total working area of the radio jet , i.e. the primary hotspot . the hotspot b1 is unresolved in our 8 ghz maps , i.e. smaller than 1.5 kpc .
the size of primary hotspots in powerful radio sources is typically 1 kpc ( e.g. laing 1988b ) . assuming that the deflecting cloud has a diameter of @xmath84 kpc , similar to the determined for the enhanced ly@xmath2 emission region at b1 , and that it is spherical , the minimal density gives a total mass of @xmath115 m@xmath116 .
the ly@xmath2 flux of the region of enhanced emission at the position of b1 indicates the presence of @xmath57 m@xmath116 of ionized gas ( see table 5 ) .
if the emission line clouds have dense neutral cores then the total mass of gas ( neutral and ionized ) may be an order of magnitude higher .
the ly@xmath2 image and high resolution spectroscopy show that the enhanced emission at the location of radio hotspot b1 has a velocity of @xmath97 km s@xmath5 with respect to the main ly@xmath2 gas . to determine whether the interaction with the radio jet could have accelerated the gas we have to estimate the jet power of 1243 + 036 .
a rough estimate can be made from the total radio luminosity ( see table 3 ) and the conversion efficiency factor @xmath163 , l@xmath164 ( e.g. bridle & perley 1984 ) , where the efficiency is probably of order 0.10.01 . thus a jet power @xmath165 ergs s@xmath5 is a reasonable assumption .
alternatively we can estimate the jet power , assuming that the total minimum energy of the radio components is replenished in the synchrotron lifetime , giving l@xmath166 ergs s@xmath5 .
a third estimate comes from the jet thrust of a relativistic jet ( not to be confused with the hot spot advance speed which is only @xmath167 ) , @xmath168=l@xmath169 ) ( @xcite ) , that can be determined from the minimum pressure and area of a hotspot . assuming a hotspot size of 1 kpc ( see above ) and jet parameters like in cygnus a , @xmath170 , @xmath171 ( e.g. perley et al . 1984 )
, we find l@xmath172x10@xmath173 ergs s@xmath5 .
thus it is reasonable to assume that the jet power of 1243 + 036 is @xmath165 ergs s@xmath5 . a gas mass of 10@xmath144 m@xmath116 with a velocity of 1100 km s@xmath5 has a kinetic energy of 1.2x10@xmath174 ergs .
the jet can have accelerated this mass over a period of 10@xmath156 years , if the transfer efficiency of jet power to kinetic energy of the gas is @xmath81% . in this deflection scenario
the pre - existing massive cloud that is hit by the radio jet does not necessarily have to be one kiloparsec - scale cloud , which given the densities of the emission line gas would be jeans unstable , but could consist of a region of small dense clouds or filaments and low density gas in between . if the clouds are optically thick they can contain large amounts of neutral gas .
whatever bends the radio jet has to be a continuous surface over the working area of the radio jet .
thus , if it is the emission line gas that deflects the radio jet , the clouds must have been turned into a `` mist '' .
the interaction with the radio jet may rip apart dense neutral cloud cores forming a gaseous medium with an increased filling factor and covering fraction ( @xcite ) .
this `` mist '' could be the medium in which the de laval nozzle forms by which the radio jet is delfected the larger filling factor will cause a larger fraction of the gas to be exposed to the ionizing radiation .
this will enhance the lyman @xmath2 emission from the region .
local shock ionization due to the impact of the jet may further enhance the ly@xmath2 emission . alternatively icke ( 1991 )
showed that a jet can be bent smoothly by a pressure gradient and thus may bounce off a gas cloud . in the case of 1243 + 036
the redirected jet would then have several emission knots due to weak shocks from further interactions with the emission line gas .
although this might account for the observed radio structure of 1243 + 036 and the acceleration of emission line gas at b1 , it is not clear if this mechanism would also cause the enhanced ly@xmath2 emission from the deflecting cloud .
\(ii ) _ precessing radio jet . _ an alternative explanation for the bent radio structure with multiple hotspots might be precession of the radio jet .
the direction of the southern jet and the slight northern extension of radio lobe a give the radio source an overall z shape , suggesting that the jet may be rotating .
as the tip of the jet moves over the wall of the radio lobe it would create new hotspots , as in the `` dentist s drill '' model of scheuer ( 1982 )
. the spectral steepening towards the outer parts of the radio structure could be the result of synchrotron aging in older hot spots .
but it could also result from there being weaker shocks with distance along the jet in the deflected flow model .
our crude estimate of the synchrotron age of hotspot b4 of 10,000 years ( see table 3 ) indicates that if the multiple hotspots are due to the directional change of the radio jet , for b4 still to be visible when b1 is being formed , the jet precession speed would be 0.001@xmath26 yr@xmath5 and the jet would have revolved 27 times within the age of the radio source ( @xmath810@xmath143 yr ) .
the projected velocity of the jet impact site from b4 to b1 ( 10 kpc ) would be @xmath175 . at these precession speeds
we would not expect the jet to be straight from the core - jet component to b1 ( and a1 ) and a straight line of hotspots from b1 to b4 , but a more gradually bent s shape structure .
also at this precession speed the time that the jet feeds the area of one hotspot is a factor 1001000 shorter than the time required to build up the hotspot ( see above ) . nor
would it have time to accelerate the emission line gas at b1 to its observed velocity .
even if we have underestimated the age of b4 by a factor of 10 , the inferred speeds seem implausibly high .
thus we regard the idea of a precessing jet causing the bent radio structure and multiple hotspots as very unlikely .
also the magnetic field projecting along the southern jet ( fig .
7 ) is consistent with flux freezing and shear flow along the jet ( @xcite ) , and hence supports the idea of a deflected flow model .
we see the bent radio structure projected on the plane of the sky . however , the bending of the southern jet may have taken place partly into the direction of the line of sight .
if the jet is still moving relativistically after the deflection , doppler boosting may increase the observed flux of weak radio knots along the jet . in the light of the orientation unification scheme for radio galaxies and quasars , 1243 + 036
would be oriented close to the plane of the sky .
if the bending of the jet towards us is as much as it is projected on the plane of the sky ( @xmath176 degrees ) , the southern jet could have been deflected to as close as @xmath830 degrees from our line of sight . for a relativistically moving jet ( @xmath177 a few )
the observed flux could have been doppler boosted by a factor @xmath8315 relative to its original direction .
the synchrotron emission associated with weak shocks in the deflected southern jet may have been enhanced forming the observed knots b2 , b3 and b4 .
\(iii ) _ sweeping of the radio jet by the ambient gas . _ as a third possibility , the sudden bend and overall z shape of the radio source may be due to `` sweeping '' of the radio jet by the ambient medium after the jet has been slowed down in the hotspots b1 and a1 .
such a mechanism has also been invoked to explain the linear radio features in some seyferts and the z shape of the low redshift radio galaxy 3c293 ( @xcite ) where the radio jets are propagating through a rotating gaseous disk .
1243 + 036 shows both morphological and kinematical similarities to 3c293 and 3c305 ( @xcite ) where there is evidence for rotation and the radio morphology is z shaped . at high redshifts not only 1243 + 036 , but also 4c41.17 exhibits a radio morphology where the axis defined by the outermost radio components is rotated with respect to the axis defined by the inner components .
these morphologies suggest that large scale motion of the ambient medium may have deformed the outer parts of the radio source .
the possible rotating gaseous disk of the outer halo in 1243 + 036 indicates that large scale organized motion is present that may provide the transverse ram pressure ( sweeping effect ) that is required to bend the radio source to its observed morphology .
the velocity shear of the outer halo is consistent with solid - body rotation that , if it is sweeping the radio plasma , would indeed cause the straight radio structure of the `` southern jet '' in 1243 + 036 . scaling the calculations of van breugel et al .
( 1984 ) for 3c293 to the much larger jet kinetic energy of 1243 + 036 ( derived above ) , requires an average density @xmath178 cm@xmath113 , where @xmath163 is the efficiency of order 0.10.01 as mentioned above , to bend the jet over a distance comparable to the jet diameter .
thus , the jet must have lost more than 99.9% of its kinetic energy in the hotspot b1 ( and a1 ) , and continue onwards as a light , but still fast jet ( @xmath179 km s@xmath5 , see van breugel et al .
1984 ) , to be bent by the ( @xmath126 cm@xmath113 ) rotating gas of the outer halo .
although we can not exclude this possibility , it seems more plausible that , with a strong interaction at b1 , the jet would be deflected along the lines of scenario ( i ) than that it would continue at high velocity , but with much less thrust , in the same direction .
thus , the bending of the radio structure in 1243 + 036 coinciding with accelerated emission line gas is evidence that the radio jet is deflected by the interaction of the jet with the emission line gas .
this supports the idea of barthel & miley ( 1988 ) that the frequent distortions of radio structures of quasars and radio galaxies at high redshifts are caused by interaction of the radio jet with gas clouds in the environment .
the @xmath1-band continuum morphology of 1243 + 036 is clearly aligned with the principal radio axis , the line connecting components a , n and b1 .
this alignment of @xmath1-band continuum emission with the radio axis is a commonly observed feature in high redshift radio galaxies ( @xcite ) .
in addition to this , 1243 + 036 also has a low surface brightness continuum component ( see fig . 8) , which has the same narrow alignment with the principal radio axis and apparently extends beyond the radio lobes .
if this faint extended continuum emission is real , this can have important implications for the interpretation of the alignment effect in high redshift radio galaxies .
one of the first proposed explanations for the `` alignment effect '' was jet induced star formation ( @xcite ) where shocks from the radio jet and the overpressured radio lobes would lead to star formation in dense clouds along the direction of the jet propagation .
daly ( 1992 ) argues that inverse compton scattering of cmb photons of the relativistic electrons in the radio plasma could cause the alignment effect at high redshifts .
optical polarization measurements of intermediate and high redshift radio galaxies have indicated that much of the aligned continuum light is polarized , therefore some or all of the light is scattered light from a hidden quasar by dust or hot electrons ( @xcite ) .
this indicates that , although jet induced star formation may play a role , scattering processes are important contributors to the alignment effect . in 1243 + 036
we see that the aligned continuum does not follow the radio structure that bends at the position of radio knot b1 into the `` southern jet '' .
instead the optical continuum extends further outwards with the same orientation as the inner continuum , the ly@xmath2 `` cone '' and the principal radio axis .
this argues against a direct relation between the optical continuum morphology and the radio jet as required by inverse compton models and jet - induced star formation and suggests that the optical continuum orientation is linked to the axis of the agn .
however , in the jet - induced star formation scenario , the extended continuum may be a relic of star formation of along a previous path of the jet , when possibly no bending occurred .
thus , the presence of faint extended optical continuum beyond the radio structure maintaining its alignment with the inner radio structure and the brightest ly@xmath2 gas favours the scenario of scattering of anisotropically emitted nuclear continuum light , although it is possible that jet induced star formation played a role at an earlier evolutionary stage of the radio source and currently in the inner parts of the radio galaxy . a much deeper image is needed to confirm the reality of the faint extended continuum emission .
we have found and studied the distant radio galaxy 1243 + 036 at @xmath0 .
it has spectacular properties both in the optical and in the radio : 1 .
the radio source extends over 55 kpc , has a sudden bend and multiple hotspot structure .
all components have ultra steep spectra and the radio emission is strongly depolarized .
2 . the radio galaxy has strong and extended ly@xmath2 and o[iii ] emission .
the ly@xmath2 emission is aligned with the radio source and shows a clumpy structure .
it is characterized by three components : ( i ) the inner halo inside the radio structure with a large velocity dispersion , ( ii ) enhanced emission with blueshifted velocity at the position of the bend in the radio structure and ( iii ) a quiescent outer halo with a velocity shear of 450 km s@xmath5 extending over a total diameter of at least 135 kpc , aligned with the inner optical and radio axis .
the @xmath1 band continuum is aligned with the principal radio axis and shows a faint extension slightly beyond the radio structure .
this emission does not have a sharp bend like the radio emission but stays aligned with the inner continuum , the emission line gas and the principal radio axis .
we have argued that the observations have the following implications : a ) the quiescent outer ly@xmath2 halo must predate the onset of the radio source .
the origin of this gas and its kinematics is probably due to the accretion of gas from the environment producing a large scale rotation , although there is a possibility that it is caused by a massive outflow from the central galaxy .
b ) the interaction of the radio jet with the emission line gas is responsible for the bent radio structure .
the jet has accelerated the gas at the interaction to a velocity of 1100 km s@xmath5 relative to the rest of the system and may have shred the emission line clouds , locally enhancing the ly@xmath2 emission .
the bending of the radio structure in 1243 + 036 coinciding with accelerated emission line gas is evidence that gas clouds may be the general cause of bending in high redshifts radio sources .
c ) the large velocity dispersion of the emission line gas inside the radio structure and the low dispersion of the gas outside of it , support the idea that the large velocity dispersions of the emission line gas in hzrgs is caused by entrainment of gas by the radio jet or shocks and turbulence in the radio plasma .
d ) the close alignment of the principal radio axis with the inner ly@xmath2 gas and especially with the ly@xmath2 from the outer halo argues for photoionization by anisotropically emitted radiation from the agn .
e ) the extended ly@xmath2 emission suggests densities of the emission line filaments of order a few to 100 cm@xmath113 .
the required confinement of the emission line gas and the radio hotspots along with strong depolarization of the radio emission argue for the presence of a dense ( 0.1 cm@xmath113 ) hot ( @xmath54 k ) halo surrounding the system .
f ) the extension of the aligned optical continuum emission beyond the radio structure and not following the bending of the radio source is an argument against inverse compton scattering and jet - induced starformation for the alignment - effect in this object .
it favours scattering of anisotropic nuclear continuum radiation .
the most likely scenario for explaining our observations is that a large rotating gaseous disk , originating from accretion associated with the formation of the galaxy , is ionized by anisotropically emitted continuum emission from the agn .
the interaction of the radio source with the gas bends the radio jet and causes large velocities in the ly@xmath2 halo as the radio plasma propagates outwards . in the outer halo
the gas is still unaffected by the radio plasma and the quiescently rotating gas disk is still intact .
because of its large luminosity and spatial extent , 1243 + 036 provides an important laboratory for investigating the intrinsic nature of high reshift radio sources and the conditions in the early universe .
further observations of this object are underway using ground based telescopes and the hubble space telescope .
we acknowledge support from an eec twinning project and a programme subsidy granted by the netherlands organization for scientific research ( nwo ) .
we thank steve eales and steve rawlings for their kindness to obtain a @xmath41 band image . | we present the discovery and detailed observations of the radio galaxy 1243 + 036 ( = 4c03.24 ) at a redshift of @xmath0 .
the radio source was selected on the basis of its extremely steep radio spectrum , suggesting that it might be very distant .
the radio source was identified with a galaxy of @xmath1 magnitude 22.5 .
subsequent spectroscopy showed strong lyman @xmath2 and [ oiii ] emission , indicating that the object is a radio galaxy at @xmath0 .
high resolution ( @xmath3 ) radio maps show an frii type radio source with a sharply bent radio structure .
strong depolarization of the radio emission indicates that the source is embedded in a magneto - ionic medium .
the most spectacular feature of 1243 + 036 is the presence of a ly@xmath2 halo of luminosity @xmath4 ergs s@xmath5 which extends over@xmath6 ( 135 kpc ) .
a @xmath7 resolution ly@xmath2 image shows that the emission line gas is aligned with main axis of the radio source and has structure down to the scale of the resolution .
high resolution spectra show that the ly@xmath2 emitting gas has complex kinematic structure .
the gas contained within the radio structure has a relatively high velocity width ( @xmath81500 km s@xmath5 fwhm ) .
the component of the ly@xmath2 emission that coincides with the bend in the radio structure is blueshifted with respect to the peak of the emission by 1100 km s@xmath5 .
there is low surface brightness ly@xmath2 emission aligned with , but extending 40 kpc beyond both sides of the radio source .
this halo has a narrow velocity width ( @xmath8250 km s@xmath5 fwhm ) and a velocity gradient of 450 km s@xmath5 over the extent of the emission .
the presence of the quiescent ly@xmath2 component aligned with the agn axis , but outside the radio source , is strong evidence that photoionization by anisotropically emitted radiation from the active nucleus is occurring .
various mechanisms for the origin and kinematics of the ly@xmath2 halo are discussed . because the halo extends beyond the radio structure with less violent and more ordered kinematics than inside the radio structure ,
we conclude that the outer halo and its kinematics must predate the radio source
. the ordered motion may be large - scale rotation caused by the accretion of gas from the environment of the radio galaxy or by a merger .
although alternatively the halo may be caused by a massive outflow , we argue that bulk inflow of the emission line gas is inconsistent with the most likely orientation of the radio source .
the large velocity - width of the ly@xmath2 gas contained within the radio source compared to that of the outer halo suggest a direct interaction of the radio source with the gas .
the spatial correlation of enhanced , blue - shifted ly@xmath2 emission and the sharp bend of the radio structure suggest that the emission line gas could have deflected the radio jet .
the impact of the jet could have accelerated the gas at this position and may have locally enhanced the ly@xmath2 emission . extended faint optical continuum emission is aligned with the principal radio axis , a phenomenon commonly observed in high redshift radio galaxies .
this emission it does not follow the bending of the radio jet , indicating that , at least in 1243 + 036 , models invoking scattering of continuum radiation from the agn as the cause of this alignment are favoured . |
cool white dwarfs showing traces of carbon , either as neutral carbon lines or molecular c@xmath0 swan bands , are collectively known as dq stars .
past analyses showed that these stars are helium rich with carbon abundances ranging from @xmath1 to @xmath2 as determined from optical or ultraviolet ( iue ) spectroscopic observations @xcite .
the presence of carbon has been successfully explained by a model in which carbon diffusing upward from the core is brought to the photosphere by the deep helium convection zone @xcite .
this model was shown to reproduce the observed carbon abundance distribution as a function of effective temperature fairly well for a helium layer thickness of @xmath3 , a value at odds with those obtained from evolutionary models of post - agb stars that predict much thicker helium layers of the order of @xmath4 . more recently , @xcite and ( * ? ? ?
* hereafter collectively referred to as blr ) presented a photometric and spectroscopic analysis of a large sample of white dwarfs aimed at improving our understanding of the chemical evolution of cool white dwarfs .
their sample includes several dq stars that were analyzed with pure - helium model atmospheres , the only models available at that time .
recently , @xcite studied the importance of including carbon in the atmospheres of the dq star procyon b. their effective temperature of @xmath5 k , derived from models including carbon , was significantly cooler than the earlier value of @xmath6 k obtained by @xcite on the basis of pure helium models .
this large difference was attributed to an increase of the he@xmath7 free - free opacity resulting from the additional free electrons provided by carbon .
the natural step following the comprehensive studies of blr is thus to include explicitly metals and molecules in the model atmospheres .
the model atmosphere code is a modified version of that described at length in @xcite , which is appropriate for pure hydrogen and pure helium atmospheric compositions , as well as mixed hydrogen and helium compositions . in order to treat in detail white dwarfs of the dq and dz spectral types
, we had to include metals and molecules in our equation - of - state and opacity calculations .
details will be provided elsewhere ( dufour et al . 2005 , in preparation ) . since in this analysis we are only concerned with dq stars showing the c@xmath0 swan bands , we need only to include helium and carbon .
our model grid covers a range of @xmath8 to 12,000 k in steps of 500 k , @xmath9 to 9.5 in steps of 0.5 dex , and @xmath10 to @xmath11 in steps of 0.5 dex .
the first sample used for this study is drawn from the blr analysis .
it consists of 12 dq stars with trigonometric parallax measurements , high signal - to - noise spectroscopy showing the carbon molecular bands , optical @xmath12 and infrared @xmath13 photometry .
the method used to fit the data is similar to that described at length in blr , with the exception that in addition to the effective temperature and the solid angle @xmath14 , we now have a third fitting parameter , the carbon abundance .
briefly , we transform the magnitudes at each bandpass into average fluxes .
the resulting energy distributions are then fitted with the model fluxes , properly averaged over the filter bandpasses , using a nonlinear least - squares method .
we begin by assuming a pure helium composition and @xmath15 to get initial estimates of @xmath16 and @xmath14 .
the solid angle is combined with the distance @xmath17 derived from the trigonometric parallax to obtain the stellar radius @xmath18 .
this radius is then converted into @xmath19 using an appropriate mass - radius relation for white dwarfs , and the procedure is repeated until a convergence in @xmath19 is reached .
next we turn to the spectroscopic observations and determine the carbon abundance by fitting the swan bands at the values of @xmath16 and @xmath19 obtained from the fit to the energy distribution .
this carbon abundance is then used to obtain new estimates of the atmospheric parameters from the energy distribution , and so forth .
we repeat the procedure typically five iterations until @xmath16 , @xmath19 , and the carbon abundance converge to a consistent photometric and spectroscopic solution .
the @xmath19 values are then converted into masses using the cooling models of @xcite .
our spectroscopic fits for the blr sample are displayed in figure [ fg : fitdqblr ] ; the photometric fits differ very little from those already shown in blr and will thus not be repeated here .
we compare in figure [ fg : shift ] the effective temperatures and masses derived from pure helium models ( e.g. blr ) with those obtained here with models including carbon ( see also weidemann , these proceedings ) . we find that our revised effective temperatures are reduced significantly with respect to the pure helium solutions , in agreement with the conclusions of @xcite for procyon b. the inclusion of carbon in the equation - of - state increases the number of free electrons and consequently the contribution of the he@xmath7 free - free opacity .
the resulting effects on the atmospheric structures can produce significant differences in the derived atmospheric parameters . since the @xmath16 values have been greatly reduced , so are the model fluxes , and larger solid angles are required to fit the photometric data , which imply larger stellar radii or smaller masses , as shown in figure [ fg : shift ] .
an interesting consequence of our findings concerns the mass distribution of dq stars . an intriguing question raised by blr was the fact that the mean mass of dq stars , @xmath20 , was much higher than that obtained for their likely progenitors , the db stars , by @xcite , @xmath21 , based on spectroscopic fits .
our improved analysis brings the mean mass of dq stars down to @xmath22 , in much better agreement with the spectroscopic mean mass of db stars , and more in line with our understanding of white dwarf evolution .
our second sample consists of 40 stars spectroscopically identified as dq ( human i d ) in the first data release of the sloan digital sky survey ( sdss ) .
details concerning these observations are described in @xcite and references therein .
prior to the sdss study , only 37 cool ( @xmath23 k ) white dwarfs with carbon features had been analyzed by various investigators , and the sdss first data release _ alone _ has more than doubled the number of dq stars known , allowing us to improve significantly the statistics of dq stars .
the method we use here to fit the sdss data is similar to that described above , with the exception that the sdss @xmath24 photometry is used instead .
these photometric passbands cover the entire optical range from the uv atmospheric cutoff ( 3200 ) to the red sensitivity cutoff of the detector ( @xmath25 ) .
furthermore , since trigonometric parallax measurements are not available for the sdss stars , we assume a value of @xmath26 for all objects .
our results are summarized in figure [ fg : pelletier ] where we show the carbon abundances as a function of @xmath16 for both the blr and sdss samples .
we see that the carbon abundances form a relatively narrow sequence in the @xmath27 plane . the sequence defined by the sdss sample overlaps the blr sequence quite nicely , suggesting that the physical characteristics of these two samples are very similar . since the blr sample as a whole has a normal mean mass near 0.6 @xmath28 , the value of @xmath29 assumed for the sdss sample is probably reasonable .
it is also comforting to see that procyon b ( star symbol at 7740 k ) , with a precise mass determination of 0.602 @xmath30 determined by @xcite , falls well within our sequence .
our results are thus entirely consistent with the idea that all dq stars evolved from hotter db white dwarfs . also shown in figure [ fg : pelletier ] are the results from two different sets of evolutionary model calculations , which predict the carbon abundances as a function of decreasing effective temperatures for various values of the thickness of the helium layer @xmath31 .
the dashed curves represent the calculations of @xcite at 0.6 @xmath28 with , from top to bottom , @xmath32 , @xmath33 , and @xmath34 . as discussed in the introduction , our abundance determinations would suggest values of @xmath35 smaller than @xmath33 for all dq stars , much thinner than the values predicted by post - agb evolutionary models .
a new generation of evolutionary models computed by fontaine & brassard ( these proceedings ) yield quite different results , however , as shown in figure [ fg : pelletier ] by the solid curves for models at 0.6 @xmath28 with , from top to bottom , @xmath32 , @xmath34 , and @xmath11 . with these new models ,
the observed carbon abundances in dq stars are consistent with values of the helium layer thickness between @xmath36 and @xmath2 , in much better agreement with the results from post - agb models .
several objects in figure [ fg : pelletier ] show carbon abundances that lie about 1 dex above the bulk of dq stars . among these
, the only white dwarf with a measured trigonometric parallax is g47 - 18 ( wd 0856 + 331 ; top filled circle ) with an estimated mass of @xmath37 , by far the most massive dq star in figure [ fg : fitdqblr ] .
it is thus tantalizing to suggest that the dq stars with larger - than - average carbon abundances represent the high mass component of the white dwarf mass distribution .
@xcite has suggested that the hot dq stars discovered in the sdss could all be massive , and could correspond to the missing high mass tail of the db star mass distribution @xcite .
if both hypotheses are confirmed , the cool dq stars with large carbon abundances observed here would simply represent the natural extension of the hot dq stars identified by liebert et al .
finally , the absence of neutral carbon lines in the optical spectra of db stars with normal masses is naturally explained by our results .
indeed , the expected carbon abundances in hot ( @xmath38 k ) helium - rich white dwarfs with @xmath35 between @xmath39 and @xmath2 fall below the detection threshold of the atomic c i lines according to figure [ fg : pelletier ] . | we present an analysis of spectroscopic and photometric data for cool dq white dwarfs based on improved model atmosphere calculations .
in particular , we revise the atmospheric parameters of the trigonometric parallax sample of @xcite , and discuss the astrophysical implications on the temperature scale and mean mass , as well as the chemical evolution of these stars .
we also analyze 40 new dq stars discovered in the first data release of the sloan digital sky survey . |
the sharp flux change is characteristic of the spectral split phenomenon @xcite . ]
neutrino flavor conversion in core - collapse supernovae has revealed to be complex .
three features produce characteristic flavour conversion phenomena : the presence of large neutrino number densities , of ( front and reverse ) shock waves @xcite and of turbulence @xcite .
pantaleone first pointed out the importance of the neutrino - neutrino interaction and the fact that it would introduce a non - linear refractive index @xcite . at the mean - field level
the effective hamiltonian includes a term which is proportional to the neutrino density matrix ( instead of the matter number density as in the msw effect ) to encode the fact that neutrinos can change their flavour from the neutrinosphere to the point where they interact with another neutrino .
it was recognized in @xcite and recently in @xcite that the flavour evolution is drastically modified after the implementation of such a term .
three flavour conversion regimes might be present in nature .
they consist first in a collective synchronization of the neutrinos with no flavor conversion , then in the occurrence of `` bipolar '' flavour oscillations and finally in a complete swap of the ( anti ) neutrino spectra above a critical energy - the spectral split phenomenon ( figure 1 ) ( for a review see @xcite ) . a useful formalism to pin down the phenomena underlying the flavour conversion effects , observed in numerical calculations , is offered by the quantum mechanical analogy with effective spins .
this consists in associating to the neutrino amplitudes an effective spin @xmath10 , to the hamiltonian an effective magnetic field @xmath11 , and solve a precession equation instead of the schrdinger equation @xcite . in particular , the third component of the spin ( also named isospin ) gives the neutrino survival probability , while the @xmath12- and @xmath13-components depend upon the neutrino mixings .
this analogy was already used to describe vacuum oscillations as well as the msw effect ( see e.g. @xcite ) .
( note that , in these cases , the effective magnetic field only has real components . )
the msw resonance occurs when the @xmath14-component of the magnetic field becomes zero .
an adiabatic evolution corresponds to the fact that the spin precession frequency is fast compared to the magnetic field evolution , so that the spin evolves with it through the resonance instead of lagging behind ( non - adiabatic solution ) .
the @xmath15 interaction term has complex contributions because of its dependence upon the neutrino density matrix .
this introduces a time - varying effective magnetic field in the plane .
ref.@xcite has shown that in the synchronization regime the total @xmath10 precesses around @xmath11 ( almost aligned with the @xmath14-axis ) , dominated by the @xmath15 interaction , so that no flavour conversion occurs . using the analogy with a pendulum , in @xcite
it has been pointed out that , contrarily e.g. to the well known msw effect , collective flavour conversion effects occur for any value of the third neutrino mixing angle , while such a parameter needs to be stricly non - zero .
bipolar oscillations can be associated to the evolution of a flavour pendulum @xcite and a gyroscopic pendulum @xcite that makes both a precession and a nutation .
this picture is further investigated in @xcite . concerning the spectral split phenomenon associated with a sharp spectral swapping above a critical energy ( figure 1 ) , in @xcite the authors make the hypothesis that at the end of the synchronization and of the bipolar regimes , the neutrino evolution follows a collective precession mode until neutrino densities are low and at the late stage of this precession solution the stepwise swapping occurs .
the existence of a comoving frame is pointed out in the early work @xcite . in @xcite ,
an explicit adiabatic solution is postulated in the comoving frame an msw - like solution that is shown to behave like the spectral split solution .
the split energy is explained by the ( approximate ) conservation of the total neutrino lepton number ( for the two - flavour case ) @xcite .
the extension to three flavours , based upon su(3 ) is first studied in @xcite .
ref.@xcite shows that there exists an exact solution to the full two - neutrino problem ( in the single - angle approximation and if the matter term is neglected ) .
the authors put this neutrino hamiltonian in relation with the class of gaudin models and , in particular , the ( reduced ) bcs pairing hamiltonian describing superconductivity .
the spectral split phenomenon is understood in terms of the adiabatic evolution from the quasiparticle degrees of freedom in the high - density region where they coincide with the flavor eigenstates , to the vacuum , where they coincide with the mass eigenstates . , @xmath16 .
these coincide with the place and the neutrino energies that undergo the spectral split phenomenon @xcite .
( note that , when the solid line upturns , the conditions are not both fulfilled , so that no spin flip - or flavour change - occurs ) . ] in a recent work , we have been showing that the spectral split phenomenon is a magnetic resonance ( mr ) phenomenon @xcite . in its simplest realization , the mr consists in a spin - flip occurring in presence of a constant and a time varying magnetic fields .
figure 2 shows the result of a two - flavour numerical calculation that fully includes the mixings , the matter , the neutrino - neutrino term and uses a realistic supernova density profile .
the solid line corresponds to the neutrino energies and the location where the magnetic resonance conditions are fulfilled , namely @xmath17 , @xmath16 , that coincide with the neutrino energies and the location in the supernova , where the spectral split phenomenon occurs .
note that , saying that a mr phenomenon occurs in particular implies the existence of an msw - like solution in the comoving frame .
this frame has been numerically identified ( in the matter basis ) , as given by the average of the ( fast varying ) magnetic field .
the results of ref.@xcite give a consistent picture , in the flavour and in the matter basis , of the neutrino flavour evolution during the spectral split phenomenon .
an important aspect , emerged with the @xmath15 interaction inclusion in the neutrino flavour evolution simulations is the emergence of instabilities in flavour space . refs.@xcite have shown that bipolar oscillations start because of the presence of an instability . indeed
, such instabilities appear under various conditions , and can be present for any neutrino mass hierarchy , depending on the primary neutrino fluxes @xcite ( multiple spectral splits ) , or in multi - angle calculations @xcite .
a linearized analysis of instabilities is proposed in @xcite .
ref.@xcite has shown that the instability associated with the start of bipolar oscillations is due to a rapid growth of the matter phase .
a key question for observations is to determine how much of the flavour conversion effects described survive , when inputs ( neutrino fluxes , neutrinospheres , matter density profiles ) from multi - dimensional ( exploding ) supernova simulations are employed .
for example , the presence of matter can introduce decoherence ( see e.g. @xcite ) and , eventually suppress these effects @xcite , during some phases of the explosion .
one should be aware that flavour conversion phenomena induced by the @xmath15 interaction occur in a region , deep in the star , while the msw resonances , produced by the neutrino interaction with matter , occur in the low density layers .
this fact is particularly intriguing because on one hand such effects can be relevant for the explosion mechanism ( depending upon their exact location in realistic three - dimensional simulations ) , and for nucleosynthesis .
in fact , the significant impact on the electron fraction is pointed out in ref.@xcite while ref.@xcite has shown that predicting the @xmath18-process abundance require accurate descriptions of the @xmath15 effects . for example
ref.@xcite has explored a possible impact on the supernova explosion and found that , in principle , these conversion effects occur outside the stalled shock enhancing heating by only a few percent . however answering
this key issue might require relaxing some of the approximations made in flavour conversion studies and realize demanding numerical calculations @xcite .
much more work is definitely required to quantify the impact of flavour conversion on these open issues .
one motivation to study and observe core - collapse supernova neutrinos is the possibility to learn more about key unknown neutrino properties .
the neutrino mass hierarchy problem arises from the fact one of the sign is unknown : the third mass eigenstate can be the heaviest , or the lightest .
the @xmath19 sign could be determined by exploiting neutrino flavour conversion in matter in complementary ways , i.e. either in long - baseline accelerator experiments , or by measuring astrophysical neutrinos .
one possibility is the observation of core - collapse supernova neutrinos produced in future explosions ( see e.g. @xcite ) , or of athmospheric neutrinos going through the earth and detected in ice or water cherenkov detectors ( see e.g. @xcite ) .
more precisely , it is shown in @xcite that , if the hierarchy is inverted , a dip or bump should be present in the positron time signal associated with a supernova explosion , depending upon the impinging neutrino energy , in a cherenkov or scintillator detector .
the same applies to the neutrino signal , if the hierarchy is normal .
the value of the cp violating dirac phase is one of the key unknowns that will be among the goals of the future neutrino experimental searches . if non - zero
, it can have an impact in astrophysical and cosmological environments ( besides leptogenesis ) .
@xcite first studied cp effects in the supernova context ( with a negative result ) .
the existence of possible cp violating effects in supernovae has been established in @xcite and @xcite in presence of the neutrino - neutrino interaction . in particular , effects can arise due to loop corrections or to physics beyond the standard model , or to any physics breaking the factorization condition established in @xcite .
a first quantification of the dirac phase impact on the neutrino fluxes shows modifications at the level of @xmath20 .
it is still too early to exclude that an enhancement of such effects could be present and , maybe , have an impact on observations .
ref.@xcite has studied for the first time the possible effect of a non - zero dirac phase on the neutrino degeneracy parameter and on the primordial light element abundances . by considering three active flavours only ,
modifications at the level of 1@xmath21 of the primordial @xmath22he abundance , due to a non - zero dirac phase , are found . in conclusion ,
the domain of supernova neutrinos is steadily progressing .
exciting novel flavour conversion phenomena have been uncovered .
their understanding and interplay with realistic supernova modelling and nucleosynthetic simulations still require serious efforts .
mantaining a network of neutrino detectors , and possibly constructing a large - size one , might offer the unique opportunity to pin down the secrets of supernovae .
y. fukuda _ et al .
_ [ super - kamiokande collaboration ] , _ phys . rev .
* 81 * , 1562 ( 1998 ) [ hep - ex/9807003 ] .
k. abe _ et al .
_ [ t2k collaboration ] , _ phys . rev . lett .
_ * 107 * , 041801 ( 2011 ) [ arxiv:1106.2822 ] . y. abe _ et al . _
[ double - chooz collaboration ] , _ phys .
* 108 * , 131801 ( 2012 ) [ arxiv:1112.6353 ] .
f. p. an _ et al . _ [ daya - bay collaboration ] , _ phys .
* 108 * , 171803 ( 2012 ) [ arxiv:1203.1669 ] .
j. k. ahn _
[ reno collaboration ] , _ phys .
lett . _ * 108 * , 191802 ( 2012 ) [ arxiv:1204.0626 ] .
j. m. conrad and m. h. shaevitz , _ phys .
lett . _ * 104 * , 141802 ( 2010 ) [ arxiv:0912.4079 ] . c. volpe , _ j. phys .
g _ * 34 * , r1 ( 2007 ) [ hep - ph/0605033 ] .
s. ando , j. f. beacom and h. yuksel , _ phys .
lett . _ * 95 * , 171101 ( 2005 ) [ astro - ph/0503321 ] .
d. autiero _ et al .
_ , _ jcap _ * 0711 * ( 2007 ) 011 [ arxiv:0705.0116 ] .
d. vnnen and c. volpe , _ jcap _ * 1110 * , 019 ( 2011 ) [ arxiv:1105.6225 ] .
f. t. avignone and yu .
v. efremenko , _ j. phys .
g _ * 29 * ( 2003 ) 2615 .
r. lazauskas and c. volpe , _ j. phys .
g _ * 37 * , 125101 ( 2010 ) [ arxiv:1004.0310 ] . c. volpe , _ j. phys .
g _ * 30 * ( 2004 ) l1 [ arxiv : hep - ph/0303222 ] .
j. t. pantaleone , _ phys .
b _ * 287 * , 128 ( 1992 ) .
s. samuel , _ phys .
d _ , * 48 * , 1462 ( 1993 ) . r. f. sawyer , _ phys
. rev .
d _ * 72 * , 045003 ( 2005 ) [ hep - ph/0503013 ] .
h. duan , g. m. fuller and y. z. qian , _ phys .
d _ * 74 * , 123004 ( 2006 ) [ arxiv : astro - ph/0511275 ] . h. duan , g. m. fuller , j. carlson and y. z. qian , _ phys .
d _ * 74 * 105014 ( 2006 ) [ arxiv : astro - ph/0606616 ] .
h. duan , g. m. fuller and y. -z .
qian , _ ann .
nucl . part .
* 60 * , 569 ( 2010 ) [ arxiv:1001.2799 ] .
s. hannestad , g. g. raffelt , g. sigl and y. y. y. wong , _ phys .
d _ * 74 * , 105010 ( 2006 ) [ erratum - ibid .
d * 76 * , 029901 ( 2007 ) ] [ arxiv : astro - ph/0608695 ] .
h. duan , g. m. fuller , j. carlson and y. -z .
qian , _ phys .
d _ * 75 * , 125005 ( 2007 ) [ astro - ph/0703776 ] .
m. -r .
wu and y. -z .
qian , _ phys .
d _ * 84 * , 045009 ( 2011 ) [ arxiv:1105.2068 ] .
g. g. raffelt and a. y. smirnov , _ phys .
d _ * 76 * , 081301 ( 2007 ) [ erratum - ibid .
d * 77 * , 029903 ( 2008 ) ] [ arxiv:0705.1830 ] . | we summarize the recent developments in our understanding of neutrino flavour conversion in core - collapse supernovae and discuss open questions .
address = institut de physique nuclaire orsay , cnrs / in2p3 , f-91406 orsay cedex , france the existence of neutrino mixings and non - zero neutrino masses is responsible for the neutrino oscillation phenomenon
these elusive particles change their flavour when they travel , as first discovered by super - kamiokande @xcite .
theoretically , accounting for neutrino oscillations requires the introduction of a unitary matrix relating the neutrino flavour to the mass basis , the maki - sakata - nakagawa - pontecorvo ( mnsp ) matrix , analogue to the cabibbo - kobayashi - maskawa matrix in the quark sector . if three neutrino families are assumed , the mnsp matrix depends upon three mixing angles and three ( two majorana and one dirac ) phases .
the former are now measured with good precision , the best being the third neutrino mixing angle determined by the t2k @xcite , double - chooz @xcite , daya - bay @xcite and reno @xcite experiments .
an important open question is the value of the phases , since non - zero dirac or majorana phases render the mnsp matrix complex and introduce cp violation in the lepton sector . in particular ,
the t2k and no@xmath0a experiments are capable of covering a ( small ) fraction of the dirac phase values that might indeed include the true value .
if this is not the case , the investigation of a large fraction of the dirac phase values might ask for the construction of large size detectors such as hyper - k , or experiments like daedalus @xcite .
long term facilities , that have been discussed , include beta - beams @xcite and neutrino factories .
neutrino oscillations in vacuum is an interference phenomenon among the mass ( or propagation ) eigenstates that acquire different phases , while they travel from the neutrino source to a detector .
however , the presence of matter modifies the way neutrinos change their flavour in a significant way .
note that the neutrino interaction with the matter composing a medium is usually accounted for , at the mean - field level , by implementing an effective term to the neutrino hamiltonian .
such contribution is linear in the weak coupling constant and proportional to the matter number density .
its inclusion produces a resonant flavour conversion phenomenon known as the mikheev - smirnov - wolfenstein ( msw ) effect @xcite .
the occurrence of the resonance and the efficiency of the flavour conversion depend upon the adiabaticity of the evolution at the resonance , which in turn is determined by the specific matter density profile of the medium and the neutrino properties ( energies , mixing angles , mass - squared differences and their sign ) .
in particular , it is now established that the msw effect is at the origin of the `` solar neutrino deficit problem '' ( see e.g. @xcite ) first observed by r. davis pioneering experiment @xcite . in particular , the occurrence of an msw resonance for electron neutrinos in the sun tells us one of the mass - squared differences sign ( @xmath1 ) . at present ,
the ensemble of the neutrino oscillation experimental results is well described within the three neutrino oscillation framework .
however , some anomalies have been identified , in particular , in the reactor data ( the `` reactor '' anomaly ) @xcite and in a measurement with a static source ( the `` gallium '' anomaly ) @xcite .
the results of the accelerator - based miniboone experiment also present puzzling aspects that still need to be clarified @xcite .
if confirmed , the interpretation of these anomalies will require going beyond the employed theoretical framework and implementing , for example , the inclusion of a fourth sterile neutrino , or of some other exotic possibility .
numerous projects have been proposed to address the hypothesis of a fourth sterile neutrino having a large squared - mass difference , in the ev@xmath2 range .
these comprise experiments using either a detector located at short baselines ( e.g. @xcite ) , or a static source inside already existing neutrino detectors such as borexino or kamland as in @xcite , or @xmath3-decaying nuclei produced at radioactive ion beam facilities and impinged on a target inside a @xmath4 detector @xcite . clarifying the origin of these anomalies constitutes one of key open issues .
other crucial open questions include the determination of the neutrino mass scale , of the neutrino ( dirac or majorana ) nature and of the neutrino mass hierarchy .
the katrin experiment @xcite is about to explore the sub - ev range for the neutrino effective mass ; while double - beta decay measurements are also sensitive to the nature .
understanding how neutrinos modify their flavour in media is a fascinating theoretical problem , also crucial for observations that range from solar and core - collapse supernova neutrino experiments , to both stellar and primordial nucleosynthesis abundances .
a variety of mechanisms can be engendered , depending upon the specific medium neutrinos are traversing - a smooth matter density profile as the one in our sun , the abrupt matter change present in the earth , the turbulent media of violent events as core - collapse supernovae , the accretion - disks around black - holes ( ad - bh ) , and expanding environments as the early universe . here
we will focus upon the core - collapse supernova case and describe flavour conversion effects associated with the fact that neutrinos meet lots of neutrinos when traversing these massive stars .
core - collapse supernovae occur at the end of the life of massive stars .
while our sun continuously emits electron neutrinos , core - collapse supernovae produce neutrinos and anti - neutrinos of all flavours in a burst lasting about 10 seconds .
such explosions produce an enormous amount of neutrinos since @xmath5 of their gravitational energy is taken by about 10@xmath6 neutrinos with 10 mev average energy .
these expectations have been roughly confirmed by the electron anti - neutrino events , measured during the explosion of the sn1987a , located in the large magellanic cloud at 50 kpc ( see e.g. @xcite and references therein ) . theoretically , a significant step forward has been done , with the increase in complexity of core - collapse supernova simulations , that are nowadays based upon multidimensional computations and include convection , realistic neutrino transport and dynamical instabilities - the standing accretion shock instability ( see e.g. @xcite ) .
although a consensus is not yet reached on the relative importance of these different aspects and on the precise mechanism , different groups are now capable of obtaining succesfull explosions for various progenitor masses .
this progress makes us hope that an answer to the longstanding problem of understanding how these massive stars explode , might find an answer in the coming decade .
clearly , key information on the explosion mechanism and on neutrino properties could be gained if novel experimental data are obtained .
core - collapse supernovae in our galaxy are rare events , they become as frequent as abount one event per year at distances as large as 1 - 4 mpc @xcite ( obviously only a bunch of events could be measured from far events ) .
the network of running neutrino detectors around the world can measure a supernova explosion if it occurs in our galaxy .
if a supernova explodes tomorrow , running detectors can collect several thousands events altogether ; while if one of the large ( megaton ) size detectors under study @xcite is built , up to @xmath7 events can be collected .
so many events would give us the picture of the explosion , seen with neutrinos : their emission closely follows the different phases under which a massive star dies . in order to extract as much as information as possible on neutrino properties and the supernova dynamics
, one needs a network of detectors based upon various technologies , and having different thresholds .
ref.@xcite shows examples of how the presence of one- and two - neutron emission thresholds in a lead - based supernova detector like halo can significantly reduce the number of degeneracies .
supernova neutrino observatories exploiting nuclei to detect neutrinos , such as halo , would benefit from the realization of low energy neutrino scattering measurements of neutrino - nucleus cross sections at spallation sources @xcite , or at a low energy beta - beam @xcite .
the diffuse supernova neutrino background , produced by explosions at different redshifts , encodes information on the star formation rate as well .
studies are ongoing to determine if an upgraded neutrino detector technology , based upon the addition of gadolinium to a water cherenkow detector , allows to reach the required background suppression ( see @xcite for a review ) .
predictions indicate that large size detectors should have the sensitivity for such a discovery .
for example , in ten years one expects on average about 340 events in 440 kton water cherenkov detector , about 60 events in 100 kton liquid argon and about 100 events in 50 kton liquid scintillator @xcite . finally , neutrinos play a role in stellar nucleosynthesis processes ( neutrino nucleosynthesis , @xmath8process , @xmath9 process ) .
although such processes strongly depend upon the specific astrophysical conditions and site , as well as nuclear physics inputs , they are often sensitive to neutrino properties , such as the presence of a fourth sterile neutrino @xcite . in conclusion , accurate predictions of the supernova neutrino signal , of the diffuse supernova neutrino background and of nucleosynthesis abundances require a precise understanding of how neutrino change their flavour in these explosive environments . this domain is being actively investigated since many novel flavour conversion phenomena have been identified , while important open questions remain . |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.